Improved centering with celestron mounts

I have been doing some plate solving and centering with a cge-pro and realized there is a limitation to the accuracy that can be improved using a feature of the mount.

I found that repeated attempts to solve and center the image yielded no improvement on the arc-minute scale of error - and in fact there is no expected improvement since all it is doing is solving/syncing/slew. As long as the solve and sync are accurate, the main error is in how accurately the mount lands where it wants to when it does a slew - and that won’t be perfect in the first place because it requires timing and dead-reckoning to have the mount land at the right encoder values at the right time.

But I did find improved accuracy by using the “calibrate goto” feature of the mount - which attempts to measure the error in the goto timing and correct for it. I am imaging at 0.57" per pixel and I had set a centering accuracy of 100 pixels - and it wasn’t able to reach that. But after doing calibrate goto, it is now around 60-80 pixels - or about 1/2 arc-minute. I need accuracy because I am using OAG with a fairly small sensor (qhy5l-ii) and I want the guidestar to appear after I do the centering.

So - this is a little-used feature of celestron mounts - but it has an application here to improve accuracy in a slew. Ideally I would like an option not to do a slew to recenter - and instead to do a calculated nudge. I assume that is already a feature request. Done properly that should actually converge on the requested target location.


1 Like

You mean “Utilities” > “Calibrate Mount” > “Goto”?

I’ve not actually tried this yet. For comparison/reference, I do OK with slew/point/plate-solve/correct in Sequence Generator Pro but my image scale at 1.14"/pixel is half as fine as yours with an AT8RC + STL-11000M on a CGE Pro (I want to get as much of the hardware and software chain debugged with this more forgiving 1625mm focal length before I put the EdgeHD 1400 up on the mount and move it to a permanent structure).

I’ve had it on my TODO list for an evening when I have some spare time to play around. I need to devote an evening to meridian flip testing / practice, so I might stick the Calibrate Mount > Goto exercise into that.


Yes - that is the one. It only takes a few seconds to perform and you can do it in daylight with the camera off. You just want the mount loaded with the actual imaging setup so it can measure the system behavior with the actual weight.

I have only tried this in a single before/after session and it seemed to improve centering - so more data would be good. But Calibrate Mount -> GoTo is intended to do exactly the thing needed here for centering to work well - so it really should help. But I have sent a separate inquiry to see exactly what the expected GoTo accuracy should be after doing the procedure. 1/2 arc-minute may be about as good as it can do - and if so you would need to make sure the centering tolerance isn’t too tight - or it simply won’t get there because it doesn’t have a way to converge.

Note that this “goto accuracy” has nothing to do with the overall goto accuracy - which may be more like 3-5’. This has to do with aiming for a particular RA/Dec - and actually landing where it was aiming according to its own encoder system. That’s a separate error on top of the error in modeling the mount to the sky.


Continuing the discussion from Improved centering with celestron mounts:

Oh interesting. The docs didn’t walk through the process in detail IIRC so I assumed I need to be under the sky with actual stars to measure against when calibrating. So I can actually do it indoors with the scope loaded up as usual? That’s nice to know since once reason I’ve put it off is I didn’t want to waste clear sky time which is often kind of precious around here.

Intriguing. My understanding was that this is a “do once when you put a new equipment load on the mount and otherwise forget about it” procedure. I’m assuming it’s possible to put NexRemote in Hibernate mode and exit it, and have the goto calibration settings restored when NexRemote next starts up? Or does one want to do it each time one starts NexRemote anew? I typically run my scope from SGP through NexRemote rather than using the physical hand control…



I believe that when you do the procedure once, it is stored forever until you do it again or do a firmware reset. So you can hibernate or shut it down or anything you want and it will be retained. In theory it only needs to be done once after you do a significant change to the weight.

It literally only takes a few seconds because all it does is a quick and short slew - and I assume it measures the difference between where it wanted to go and where it landed - and uses that info to correct future slews. That is exactly what is needed here in sgp.

But again - I have only used this once and I’m not sure how much it helped - or if there is something else going on to limit the accuracy during centering. But I certainly expected the sgp centering to converge on multiple tries - but on thinking about it I realized it wouldn’t if there is systematic bias even if the plate solves and syncs are perfect - and the goto calibration is designed just for the purpose of removing that bias.


As an update - it seems the goto calibration may not help much with recent firmware on celestron mounts - so it may not be something to worry about. But there will likely be some residual error in centering after a slew - and repeated solves/syncs/slews won’t improve it. So you may need to set a centering tolerance of 1’ or so - and set the repeat value to no more than two.

Those are ballpark values based on my cge-pro - and you should confirm for yourself. But when I did repeated solves/syncs/slews during the centering routine with a narrow tolerance - it quickly converged to some fixed error that did not improve.

If the centering code could do the final centering motion with calculated nudges rather than a slew - that would allow it to converge instead of going to the same offset. But 1’ may be fine for most people.

I went ahead and entered a feature request to nudge instead of slew for centering - when close to the target.