Hi Bruce, Bill,
I have the utmost respect for your skepticism with respect to this and I will try and address your doubts accordingly. This should be an interesting exercise because either there's a problem with the particular build of PHD I was using when I obtained those results or some folks are about to get a lesson on how to potentially, depending on their hardware, improve their guiding.
First of all, let me start by saying that -- in case anyone has their doubts -- I most certainly did not doctor or fabricate these results in any way. I have better things to do than to waste my own time and I have absolutely no intention of wasting yours either.
It was my honest belief at the time that those results were not that remarkable, as I have seen comments by EQ8 owners which speak of numbers in and around that ballpark. Of course, the EQ6 and EQ8 are very different beasts indeed and this can be seen by comparing the worm gear specifications which are listed in one of the EQMod documents, assuming they are accurate.
My particular mount is the Skywatcher AZ-EQ6-GT and until I made a point of checking a moment ago, I was under the impression that this mount was basically identical to the EQ6Pro/NEQ6. I see now that those mounts are not belt-driven as standard but must be modified with some sort of kit. I can't speak for how much difference the belt drive makes to the standard EQ6 drive or whether a modified EQ6 performs the same as the factory-shipped AZ-EQ6-GT belt drive. I do know that the EQ8 is belt driven as standard, which is partly to do with why I didn't find those results to be that extraordinary. I also know that the AZ-EQ6-GT has been through at least one revision since it's release, as I have owned two of these mounts and the first one was the original release.
I'd like to reiterate that this guiding camera simply doesn't suffer from hot pixels. To illustrate this, I have taken three screen captures of the camera running under PHD2 at 0.1s, 0.5s and 3s. The camera sensor is cooled to somewhere between 0°C and 1°C and in an attempt to reasonably replicate the in-field conditions, these images were captured in the dark room with the camera connected to the telescope. The camera was illuminated by shining a faint beam of infrared light (transmitting @ 935nm) parallel to but not directly at the telescope objective. Unlike the original samples I posted, I lowered the contrast to reveal any noise which would usually be hidden by having the contrast slider all the way to the right. "Use bad pixel map" is selected. I'd be interested to know how PHD could successfully complete a calibration run against a single hot pixel, since PHD would raise a "star did not move enough" exception. More importantly, if it had indeed been a hot pixel, there is absolutely no way I could have captured 18 x 10 minute subs without any actual guiding going.
The telescope is a 900mm APO refractor and those guiding results were collected using an Orion slim form-factor OAG but I have attained the same level of accuracy with an external guide scope. In the guidescope however the Atik is undersampling somewhat at 450mm.
I've attached an image of the Orion Nebula which I captured during one of those sessions. The other session involved a 4 hour exposure of the Tarantula Nebula, which I haven't attached as it does little to reflect the accuracy of the guiding that night; the image was more or less ruined after the focuser slipped when the temperature dropped to -4°C.
I'll now discuss my method for equipment setup, which will hopefully offer some insight into how different methods of preparation at the start of the night will impact guiding results later. I doubt my methods vary that much from the majority, but as you probably know, there are plenty of amateur astrophotographers who cut corners or underestimate the importance of key factors.
I don't have a permanent observatory setup, so I have to more or less assemble and configure everything from scratch prior to each session. I usually devote the first night setting up and after parking and covering the telescope, spend the entirety of the following night imaging.
- Levelling and balancing the mount. Nothing special here, since neither are particularly critical unless you have a fairly heavy load or one that is heavily biased in one direction over the other. I don't trust the bubble level on the mount, so I use a digital level and vibration dampening pads. I always balance with a very slight bias towards the focuser end in Dec and towards the OTA side on RA. On some mounts this is difficult to do properly in RA because of the tension that remains even with the clutch fully disengaged. Balancing should be done with all of the gear you intend to use attached to the telescope! Attaching a 1-2kg camera, focal reducer, filter wheel and whatever else after you've balanced your telescope is a mistake that many people make.
- Alignment. Although not important for auto-guiding it helps to know if you have cone error if you intend to use the primary telescope to polar align. I always check for cone error by doing a 3-star alignment (which allows the SynScan to calculate it) and correct it before proceeding.
- Polar Alignment. You guys already know how critical this step is but it's a subject I'm extremely vocal about, given the large number of people on the forums who don't know how to do it, don't realize how
important it is for astrophotography and are reticent to ask about it.
If you do a quick polar alignment using the feature found in the telescope hand controller -- whether it be SynScan, NexStar or whatever, and then rush off to start imaging, the auto-guiding that you're going to get is going to be barely adequate at best and more often than not, less than optimal. A lot of people are happy to just use an eyepiece, some even do their polar alignment using their finder or with the polar alignment scope. If you actually care about the quality of your guiding, none of these methods are acceptable unless you're imaging at short focal lengths.
I do my polar alignment in two steps using the camera connected to the telescope and I do it at 5400mm (4x and 2x barlow stacked). I start with the standard polar alignment tool in the Synscan controller, using the bullseye overlay in PHD to center the star. This is much easier and much more accurate than using an illuminated crosshair eyepiece. Choosing a small star and moving it out of focus until it just fits comfortably into the middle ring of the bullseye overlay allows you to center it most accurately.
This alone will not guarantee you a tight polar alignment -- as I mentioned, the hand controller utility alone is mediocre in accuracy so I then refine it via manual drift align, still at 5400mm. I do this over the course of an hour or so until the star remains in the center of the PHD bullseye overlay for as long as possible. I use a faint star available near the meridian on the opposite side of the equator from the celestial pole (in my case, the SCP). The longer the focal length, the easier it is to observe and correct any drift in the star. Making very small adjustments to the azimuth and altitude at extremely long focal lengths can be difficult on some mounts. A large CCD chip helps to prevent the star from accidentally being lost from the field of view due to over-adjustment, as do properly lubricated adjustment bolts with larger handles.
This method allows you to achieve a polar alignment with a level of accuracy which is effectively limited only by the precision at which your mount can track without assistance, excluding PE. There is no such thing as a perfect polar alignment. However, you can refine your polar alignment to a point where PHD is left with very little work to do. Doing it with a camera and doing it at the longest focal length you can manage makes all the difference and this is reflected in the original PHD screen captures I posted. This degree of guiding precision also allows for minor disturbances such as light wind whilst still keeping it below the desired 0.25".
If doubts remain then I will capture my next PHD session on video to demonstrate what I have described above.
In summary, carefully refined PA at 5400mm followed by imaging and guiding at a significantly lower focal length = accuracy.
