Precision Centering with Telescope Nudge

My Celestron CGEM controlled with CPWI cannot center my target to where I need it to be. My ZWO ASI 1600 sensor is 4656 x 3520 pixels. My SCT has a focal length of 2350mm. The image scale is 0.334 "/pixel. Plate solving (with Platesolve2) only gets me within 170 to 200 pixels of the target center coordinates. SGP’s centering uses GOTO commands to move the mount. The mount error cannot be reduced. With a 170+ pixel error for each centering, images can be off in the vertical or horizontal directions by 340+ pixels, about 10% on the vertical scale. This removes headroom in the background for many objects that I am imaging. This is a problem when imaging the same target over several nights and after a meridian flip.

I request that you add a Precision Centering capability to the centering functionality in SGP. This would use the Telescope Nudge in the North, South, East and West directions. Perhaps it could work this way:

  1. Standard plate solving is used to get the coarse centering. If within 500 pixels, then switch to Precision Centering.
  2. Since plate solving has established the error and movement directions from the first plate solve used to determine the mount’s position, nudge amounts in milliscends would be issued, much like PHD2 uses for autoguiding.
  3. Use a plate solve to determine if the nudge has moved the target to the desired center position within a user defined number of pixels.
  4. Iterate until convergence or the number of tries exceeds the user defined value.

When nudging East or South, you can take care of some of the backlash by finishing North or West. You may need a calibration iteration to determine where the nudges move the target. Dec backlash can also be a parameter be determined by the user with PHD2’s Guiding assistant.

Currently, I need to stop a sequence, and use Frame and Focus to re-position the target with the Control Panel Telescope Nudge.

Precision Centering is critical for me to automate a sequence. It would make SGP much more powerful for those of us who do not have the best equipment.

It is also important for the SCT when using autofocus. The image shifts when the primary mirror moves during focus. Precision Centering would enable the sequence to recenter on every autofocus run, or with a filter change using position offsets. That is, focus then recenter.

Mark W

One way to work round this is to use the HC and my current ASCOM driver instead of CPWI. The same problem exists in the HC but I have added code in the ASCOM driver to work round it. Reports are that the error is considerably reduced.

What you are describing involves a considerable amount of work and testing, all for very few people.


Thanks for the reply, and thank you for the updated ASCOM driver. I installed the new ASCOM driver you posted.

I prefer to use CPWI as it incorporates functionality for Celestron mounts, the StarSense accessory for alignment, ASPA for polar alignment, focus motor and gamepad integration, and most of the critical features in the HC or NexRemote application. I want to run the entire setup from my PC and not touch the HC. What I do not understand is why CPWI does not have the same accuracy as your ASCOM driver? On Tuesday, I had clear skies with good seeing. I ran the StarSense auto align, followed by and ASPA and then with a second auto align. Pointing was sufficient to get a target in the camera FOV. Plate Solving failed to get the target centered within 170 pixels. I tried Synch, Target and Scope offset modes with no accuracy improvement.

My request for Precision centering assumed that most of the functionality already exists in SGP: plate solving, telescope nudge, determination of position error and direction to go with new GOTO commands during iterative centering. The only addition is code to add the nudge amount and direction calculations and use it instead of GOTOs during the centering process.

I believe that this would be beneficial for SCT owners for fast re-centering after an autofocus or filter change. Also, there are many folks who own scopes with long focal lengths. The nudging method would be faster than issuing a GOTO and waiting for the mount to finish and settle. Also, since I use an off-axis guider with the long focal length SCT, guide stars can be hard to come by. When I do my planning with Stellarium, I set the center position for composition and the rotation for having sufficiently bright guide stars in the guide camera FOV. A significant centering error means that the guide stars can be very close to the edge of the frame.


Everything you have suggested is wrong.

Most of the functionality you have ascribed to CWPI either already exists in the HC controls or does not exist in any meaningful way.

A bit like assuming that because you know the ABC you can write Pride and Preduice, or any other literature masterpiece.

The trouble is that you have no idea of the complexity of what you are asking. I guess there is about 5,000 dollars worth of development with what you sugest. Are you prepared to pay that? Yet you are prepered to try to derail the development of SGP, deptiving everyone else of enhancements that would benefit them for your own selfish whims.

Far better to whinge at the Celestron developers, they are in a position to accomodate your desires.

Or pony up and buy a mount that you consider to be better.


Thank you for your harsh and emotional response.

I merely made a Feature Request based on my experience as well as that of other astrophographers. It is up to Jared and Ken to consider this request and to determine if it would improve SGP and what development effort is required.

Far from being selfish, I am only trying to help improve SGP. A request will not derail any development plans that Ken and Jared have in mind.

I have contacted the CPWI Celestron developers about improving pointing accuracy.

I look forward to a response from Ken or Jared.

Ken and Jared,

Recently, Celestron updated their CPWI mount control program and along with Chris Rowland’s update of the ASCOM driver, SGP plate solving in Sync mode seems to work very well. I was able to center my target to within 20 pixels (< 7") in two iterations with my CGEM II mount and a 9.25" SCT.

What I did to try to improve plate solving accuracy were:

  1. Performed two auto alignments with at least 11 points in CPWI along with good polar alignment (with Celestron’s ASPA tool).

  2. Set the location lat and long in SGP and CPWI to the same coordinates (even though plate solving should sync the scope to the solve?).

I hope this performance persists.


Mark W

I think the same sort of problem might affect more people than you would think. I’ve noticed that with my Skywatcher EQ8-R running a FL 3250 RC reduced to FL 2437, the initial slew to target is usually 2-5000 pixels off of my target. The second slew in the centering will either bring the target within 300 pixels (what I have it set to), or between 3-500 pixels, and then any further slewing by the centering action will not reduce the error at all until the centering fails and I have to start over. Because my error allowance is set to 300 pixels, that means that further nights of imaging can be up to 600 pixels off from each other, reducing the usable image. Using nudge commands after the inital goto as he described instead of slewing would allow me to reduce the allowed error and keep more of my stack instead of it being an autocrop.

I have run SGP (V3) a long time now with good plate solving performance using ASTAP. Centering a target with plate solving with synching for my Celestron CGEM II mount and a 9.25" Edge HD SCT gets me within 25 pixels (ZWO ASI 1600MM) in two or three iterations. The first slew usually has an error between 1500 and 5000 pixels. The second gets to less than 300 pixels, and the third centers properly. I believe the error is in the Celestron CPWI model and motor control. I use a Celestron StarSense accessory camera to create the sky model and alignment after polar aligning with SharpCap Pro. My CGEM performance has always been marginal and the CPWI - CGEM control may not be as accurate as the CPWI - CGX mount might be. Plate solving in SGP resolves those issues. If I leave the mount set up for multiple nights, the CPWI pointing model seems to improve with subsequent SGP plate solves, where the first plate solve can get to about a 200 pixel error. I am contemplating a mount upgrade.

Mark W