Astrophotography - Autoguiding Star Tracker Mk 2 Guiding

AUTOGUIDING WITH STAR TRACKER Mk2

Introduction

Click images to enlarge

One of my original design goals for the new tracker was to make it capable of autoguiding, as this has become an essential feature for high accuracy astrophotography. The advent of lowcost webcams, laptops and free software has made this once rare and expensive technique available to most amateurs.

Autoguiding is an advanced technique which can provide dramatic improvements to the tracking performance of a mount by using a second camera to lock onto a 'guide' star. The guide star is usually chosen from the field being imaged. A computer uses the guiding camera to detect any drift in the stars position, caused for example by poor polar alignment or drive errors. The software then provides closed loop control by feeding corrections to the drive to compensate for the errors. There are several guiding programs that are avaliable free on the web, I currently use either Guidedog or PHD Guiding software.

The images below show the autoguiding setup. For this setup the large 500mm lens and SLR take the astro-pics. The 300mm lens and webcam are used to image a guidestar. The webcam is connected to a USB port on the laptop.

A small 'red dot' sight is used as a finder to accurately point the guidescope, as finding and centering a guidestar is not easy with the small field of view the webcam provides. A second rail for mounting the finder has been fitted to the mount for the 500mm lens. This helps in accurately pointing this large lens.

Guide Scope

The webcam I use is the highly regarded Philips ToUcam PRO II (PCVC 840K). This has a CCD sensor with 640 x 480 pixels with a pixel size of 5.6 x 5.6 microns. The standard lens has to be removed and a 1.25" adaptor fitted in its place. The standard lens has an built in IR filter, which is required for accurate imaging. In its place I've fitted a Baader UV-IR rejection filter (P/No. 245 9207) to the front of the adapter, see photo opposite.

The webcam is fitted to a lens via another adaptor which I made from a Pentax 13mm extension tube and a 1.25 - 2" eyepiece adaptor.

The sensor has a small field of view and this can make it awkward to find and centre a guide star. To overcome this limitation I can either use the red dot finder described above, or fit an eyepiece to the lens to centre the guide star, then replace this with the webcam.

When the ToUcam is used with a 300mm lens the pixel resolution is 3.9 arc seconds. This is good enough for guiding a 500mm lens, (see table in Design section).

ToUcam PRO II CCD Sensor

Toucam & Pentax Lens Adaptor

ToUcam & 300 mm Pentax Lens

First Results From Guidedog - December 2006

The first set of results from Guidedog are shown below. These are for the tracker running normally without autoguiding. They show the errors in the Right Ascension (RA) tracking. The worm gear has a rotational period of 8 minutes. The chart shows two revs of the worm.

The results show:

The total tracking error is 19 arc seconds peak to peak.

Apart from the two deviations, the tracking has remained on the zero arc second line. This indicates good polar alignment and excellent tracking speed from the drive electronics. Errors from either of these sources causes RA drift away from the zero line.

The two large devitions are not sinusoidal as you would typically expect. My initial thought were that this may have been due to the slight pitting and erosion that was present on the worm due to the gyro test table spending most of its life in one particular position. (I had previously lapped the worm and gear with diamond paste to try and rectify this.)

The higher frequency components have an amplitude of approximately 2 arc sec. These errors will be from a combination motor& gearbox irregularities, software tracking errors and 'seeing'. Seeing is the effect of a stars position wobbling at high magnification due to variations in atmospheric refraction. Typically seeing can be of the order of a few arc seconds.

Investigation Of Cause of Errors

I decided to strip the tracker and investigate the possible cause of the large once per rev deviation. I used dial test indicator ( DTI ) to measure the runout on the worm shaft adjacent to the spring tensioner and found the runout profile was a reasonable match to the chart above. I removed the motor and gearbox and with the worm still tensioned against the main gear, remeasured the runout and found it had improved to 15 - 20 microns TIR.

While turning the flexible coupling by hand I noticed that it was not running true. This was caused by the Ø6.35mm hole I had drilled for the worm shaft running off from the rotational axis. (The standard hole in the coupling was Ø6mm). I concluded that when the motor and gearbox were fitted the runout of the coupling was inducing forces at the meshing interface between the worm and main gear. For one rev of the worm the forces would follow a cycle of pushing the worm into the main gear, then to one side, then forcing it to separate from the gear and then to the other side. It was when the worm was being forced away from the gear that the large deviation error was occurring. As the gears separated the contact point on the teeth would ride up causing a slowing of the tracking rate, as seen in the chart above.

The flexible coupling was not recoverable so I modified another one, achieving a better hole alignment. This coupling was fitted and aligned with the worm runout high spots and coupling high spots lined up. Using the motor to drive the worm, the runouts were remeasured and found to be the same as without the motor. The tracker was then reassembled awaiting clear skies.

Unguided Tracking Tests on Rebuilt Tracker

The chart below shows the results after the tune-up of the tracker. The blue trace shows the RA error for one rev of the worm. The downwards slope or 'RA drift' is due to poor polar alignment and is not a concern. The green trace shows the RA error with the drift removed. The tracking error has improved to ±3 arc seconds. Note: This is without any correction ie. Periodic Error Correction ( PEC ) or Autoguiding. This is extremely accurate and is an order of magnitude better than most commercial amateur telescope mounts, being on a par with some top of the range mounts, costing many thousands of pounds.

There is a very informative database of periodic errors for various mount manufacturers at Christophes Demeautis website.

Autoguiding on Rebuilt Tracker

The chart below shows the results with the autoguiding in use. After 9 minutes of running the autoguiding was switched off and the RA drift started to build up due to poor polar alignment. The autoguiding was then switched on again at 17 minutes and Guidedog quickly recaptured lock.

Although I have only just started testing out the autoguiding, I'm hugely impressed with its capabilities. Not only can it improve the tracking performance to fantastic levels, it also can be used as a tool to interrogate and tune the performance of a mount.