With some care on your part, you can coax one of these mounts to give you +/- 6 arc-seconds of accuracy during guiding (that was +/- 1 pixel for my 635mm scope + ST-8E camera at 3 arc-seconds per pixel; you may be able to do better with longer focal length scopes). In order to achieve this level of performance, you will need to devote some time to learning, documenting, and interpreting the mount's behaviors, but the reward is excellent guiding at an economical (for an astrophotographic mount) price.

The key to success with any mount used for guiding is in the setup, so one of the key questions to answer was: how effective is the mount's backlash controls? There was very minimal (basically undetectable) free play in the mount itself. However, when you install the SkySensor motors and gears, you need to take great care if you will be guiding, as it is up to you to establish a good and accurate physical connection between the motors and the mount -- a single bolt holds each motor in place. Use a straight-edge to get the gears lined up exactly in a straight line, and leave only a very small space between the gear teeth to avoid excess free play in the gear train. Don't make the gears too tight, however, or the motors will struggle! It would have been nice if the motor mounts were able to be adjusted with micrometer-like screws, but if you are careful you can get them to mesh accurately. I have included a section that covers details on gear adjustment and backlash settings.

The next step was to set the backlash for RA and Dec. I initially set these values based on testing by eye. The trick is to set RA so that there is neither a jerky start to movement in a new direction, nor any movement after releasing the button for RA movement. This should be done at around 12X or so. The trick in Dec is to find the sweet spot that gives you zero pause before motion in a new direction, but no jerkiness from too much backlash compensation.

However, when I set backlash compensation by eye this way, it wasn't accurate enough for guiding. In my initial attempts at guiding, the mount was extremely jumpy, wandering 15 or more pixels. I was worried! So I spent some time using Maxim/DL's excellent controls for moving the mount. I did lots of 1-second and 2-second movements, while making small adjustments to the backlash compensation for RA and Dec. The compensation had been set too high at first (the mount was jumping too far), and then I had it set too low (the mount would sit through numerous corrections before it actually moved). When I carefully adjusted the mount so that the 1-second movement was exactly half the distance of the 2-second movement (about 2 pixels or so), then I knew that I had zero delay (or the 1-second movement would be less than half), as well as no jerkiness (or the 1-second movement would be more than half). This method takes careful observation and an hour or more to pull off, but it's really the absolute best way to nail backlash compensation settings.

This got me within +/- 3 pixels (18 arc-seconds), so I still wasn't satisfied. At least the mount wasn't running wildly away from the guiding start point!

The next trick was to play around with the many, many settings Maxim/DL provides for mount control. After another couple of hours of testing, testing, and testing some more, I came to the conclusion that the overall goal with a mount like the GP-DX was to get the mount extremely well aligned to the pole, and to then do as little guiding as possible. This means using settings for aggressiveness, rate of movement, and duration of movement that are lower than the defaults:

• I changed the maximum move time from 3 seconds (big jump!) to .25 second. This eliminated wide excursions of the mount.
• I set the aggressiveness at 3, instead of the default 5.
• I calibrate the mount on a guide star for 15 seconds, instead of my usual 5 or 10. This tends to remove the influence of any remaining backlash in the calculated rate of movement for the mount's axes.
• I used a guide speed of .4X sidereal rate instead of the default .5X.

With these additional changes, I got the guide variations down to about one pixel plus or minus (3 arc-seconds either way, total wander 6 arc-seconds). At the focal length in use, 635mm, that was just one pixel. This rate cannot be extrapolated to longer focal lengths, since a longer focal length would also apply that longer focal length to the guide star's movements, and so the rate of increase in wander with increasing focal length is a non-linear relationship; doubling the focal length will not double the wander.

All in all, I think this establishes that the GP-DX makes a very good choice for a astro imaging. It tracks accurately enough for two-minute unguided exposures, and it guides accurately enough to have exposures as long as you need. The main thing to keep in mind is that you will need to devote some time initially to determine the optimal settings for your mount, and you will need to do a very good polar alignment each time out. The good news on polar alignment is that the polar scope is excellent, and typically provides 90-120 second unguided exposures without doing any drift alignment work.

There was almost no drift, maybe 2-3 star diameters! This is amazing to me, since I couldn't get that good of a polar alignment with my Astro-Physics 600 GTO mount using the built-in polar alignment software. I am just amazed, no other word for it.

There was slight drift, however, so I took two-minute exposures which gave me _perfectly_ round stars. A sample image of M13 shows this, as does an image of the Foxhead cluster

Both images are stacks of four or more individual images, each of which had perfectly round stars. This indicates good guiding consistency. Out of the 11 images I took of these two objects, only one had to be tossed due to star trails. The other interesting thing about these images is that they were taken through fairly heavy high clouds with a 3/4 moon high in the sky; CCD can really handle sky glow! We could see Vega, but not any of the stars in the constellation Lyra or Hercules. Extended objects like M27 were harder to image successfully over the sky glow, but I am going to stack those images and see what I managed to get.

After about an hour of slewing and tracking, I noticed that star trails started to increase, and this verifies what I suspected from the first night

Either charge your battery up really well so it is delivering 13 volts or more, or use a hefty 12V power supply. I was also successful with a 2-amp power supply, but I feel that the results were a bit more reliable with the 3-amp.

I did try to tweak the polar alignment this time, just to see if the mount was capable of handling minor adjustments, and I was able to fine-tune the alignment better than I would have expected.

The conventional wisdom has been that you need the Losmandy G-11 mount to do imaging of any quality. There have been a number of folks doing good work on lighter mounts, and it _definitely_ helps to choose a scope of a relatively short focal length (say under 800mm), or to use a focal reducer to get under that number. You will still lose an image every now and then to star movement, which I believe is due to periodic error -- about one image in ten seems to get a little streaky, and has to be discarded. Training the built-in PEC of the SkySensor will solve that problem, but losing 1 image out of 10 is not bad at all if you elect to use the GP-DX without the SkySensor.

I would definitely recommend this mount to the budding astro imager as a competent "budget" mount for CCD imaging. It costs $999 for the mount alone, and the SkySensor 2000 PC adds another $1000.

If you can get the physical backlash corrected with settings in the hand controller that do not exceed 150 units, then you have the gears in good shape. You can get this down to around 80 units if you are very, very careful, but this is hard to do as only a single bolt holds the motor and it's gear in place. Alignment to such a high degree requires a deft touch, but it can be done.

I tried calibrating the backlash visually, but I found that this was not nearly good enough for guiding purposes. I have Maxim/DL, and it has a Move button in the Guide or Focus dialog which you can use to get very good precision on the backlash settings in the hand controller. You want to be able to move the mount in + and - directions equally for equal intervals of time. Short intervals are most useful here. You may want to bump the guide speed up to .75X from .5 X for this testing. For example, put a suitable guide star on the guide chip. Now move the mount in + direction (X or Y, whichever one you want to do first) for 5-10 seconds to make sure you have taken up all the slack. Now move 1 second (or less, if you have a long focal length) in the minus direction. Note the number of pixels moved. (If there is no movement, you have not yet added sufficient backlash compensation; increase it, and begin again. A short time interval is also useful for determining if there is any remaining uncompenstated backlash.) Now move in + direction for the same time interval. The number of pixels moved should be the same. If it is not, then you must increase/decrease backlash compensation on the hand controller and retest. Continue until you get the same motions.

Repeat for other axis.

Verify lack of backlash by using a very short move interval at a longer focal length.

You may see variations in backlash compensation requirements when pointing at different areas of the sky. You can use the Maxim backlash compensation to handle these variations. I recommend that you make note of these variations, and if they are consistent and repeatable, you can use your notes to preset for various areas of the sky. I thought that the GP-DX was more prone to have these variations than other mounts I have used; perhaps my experiences were unique but if it happens to you...

By the end of this procedure, you will have a very good understanding of your mount's behaviors. If for any reason you remove the motors and remount them, you will have to repeat the procedure. This will give you excellent guiding results with the GP-DX.

The GP-DX is also very sensitive to the weight of cables and the like; if they are hanging, or are free to move very much as the scope moves, then you will get spurious variations in backlash, suddent starts and stops of the mount, or worse. The usual recommendation to run cables from the camera along the telescope to the mount, and then down to the ground, applies especially to the GP-DX. It also has fair number of protrusions which have a tendency to catch cables, as well as a few points where a cable can get trapped and pinched, to the point of cable failure!