I have seen this question asked on other groups by people who wonder whether it is worthwhile to have encoders on a mount and what it will buy you in terms of performance.
To answer the question about the utility of a high resolution encoder, I have prepared some tests to show how they enhance the performance of high res imaging on our new Mach2 mount (same applies to our larger encoder mounts). I will be showing these tests and more at the Advanced Imaging Conference in 2 weeks, but thought it might be useful to post a preview here.
Of course everyone wants round stars, but a lot of the time people forget that there is also such a thing as high resolution. I have seen plenty of images that show round stars taken on mounts that are less than adequate. The stars are round but a lot of time they are rather big, even the faint ones, which means either the focal length was short, the optics were not sharp, the seeing was atrocious or the focus was off, OR the mount was bobbling back and forth in both axes. Now, if you have a great optic and great seeing, you might want to have a mount that tracks or guides close to perfection. In imaging the mount is the most important item.
Below I have two graphs. The first one show one worm cycle (about 400 seconds) of the Mach2 with the encoders turned off. It is running with a microstep-servo and belt drive (34 tooth pulley) which people seem to think is the hot setup these days. The periodic error of the worm shows up at around 7 arc seconds. What's also shown are the considerable errors caused by the toothed belt and the inherent error of a microstep motor (about 5% for each full step). These short term errors are typically 2 - 3 times larger than what you get with our DC servo motor/spur gearbox combination, so in my opinion going just with a belt drive does not enhance performance, but it does make the mount very quiet while slewing. These errors are very common in micro-stepper belt-drive mounts that are now a dime a dozen.
The first graph shows the main periodic error, and a combination of belt and microstepper ripple of around 2 arc seconds on top of the 7 arc sec PE. About 1/2 way thru the run I gave the scope a couple of good raps with my knuckles to see how fast it would settle back. The mount is heavily loaded with almost 70lb of scope and 90 lb of counterweight. Even without the encoders operating the RA axis settled quickly to resume tracking.It lost about 2 arc seconds of position, which of course would have to be made up by a guider loop.
The second graph shows the performance with the encoders turned on. Same rap on the scope shows much faster response and lower settling time. The tracking error has dropped to below 1/4 arc second P-P with almost negligible RMS value. During the test I trimmed the tracking rate by 0.4% and then back down. The mount is controllable to any rate and drives very smoothly regardless if it's sidereal, King or custom rate, which makes it highly useful for unguided imaging using a tracking model (APPM or similar). Toward the end I sent the mount West and then back East in 0.5 arc second steps to show the response you can expect during autoguiding. If the guider sends a 1/4, 1/2, or 1 arc sec move command the mount will move by that amount, no more, no less, and with no dramatics (no overshoots, no oscillations, no retrograde motions, etc). This is what you want your mount to do. Have a look: