Spikes in Guiding after Dithering

[Tim Stevenson]

I am venturing into widefield astrophotography with a William Optics Mini Redcat 51 on a Skywatcher EQ6R Pro mount with a ASI2600MC Pro camera and need to dither and drizzle.  The guide scope is a WO Uniguide scope with a ASI120MM camera.  I am getting spikes in guiding corrections after dither... is PHD2 trying to move the mount back after the dither?  I'm asking for a 15 pixel dither.  I have attached my guidelog for a session.  Any assistance would be appreciated.

Thank you ...

[Bruce Waddington]

These dither sizes don't seem very sensible.  15 pixels at your guider image scale translates to a 90 arc-sec dither which is likely to take over 6 seconds to complete. That's too big.  I suspect you haven't thought through the difference between the image scale of your main system and the image scale of the guider.  A 15-pixel dither on the main camera might make sense but not on the guider.  Take a look at this reference:

https://openphdguiding.org/man-dev/Supplemental_Info.htm#Dithering_Ops

Regards,

Bruce

[Tim Stevenson]

Thank you Bruce ... my target dither on the main camera was 20-pixel.  It has an image scale of 4.35 and the guide scope is 5.94.  Since the guide and main scopes are fairly close in image scale, it produces a large dither request of the mount/PHD2.  If I look at my other scopes: William Optics Z81 with an image scale of 1.73 is 9-pixel, SkyWatcher Quattro 250P with an image scale of 1.03 is 3-pixel and is the same for my Astro-Tech AT8RC.  (ref: https://www.astrohowto.com/interactive-dithering-calculator/ ).  It seems all of the references I have seen (including your reference) recommend a 10 to 20-pixel dither on the main camera.

Thank you again

Tim

--
You received this message because you are subscribed to the Google Groups "Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-phd-guidi...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/cbf75179-e615-4990-9a5b-0d0a9315e140n%40googlegroups.com.

[Bruce Waddington]

These recommendations all depend on the specifics of the camera and how much movement is needed to eliminate or control fixed-pattern and walking noise.  I don't think such a coarse image scale (4.35 arc-sec/px) is typical for a main imaging set-up - something more like 1 arc-sec/px is more typical.  But if you are committed to the idea of huge dithers on the guide camera, you need to expect long settling times and large guide star excursions during the settling process.  I think you would be better advised to revisit the dither sizes and do your own testing of how much is actually required for your imaging setup.  But it's up to you.

Regards,

Bruce

[Tim Stevenson]

Thank you Bruce ... I went off and read more on dithering and drizzling.  My main objective is to drizzle due to undersampled images (as you pointed out with a coarse image scale).  My telescope is 174mm focal length and the stars can be blocky.  Drizzle helps create nice rounded stars.  One of the requirements to drizzle is to have well dithered data.  The creator of Astro Pixel Processor says that the dither only needs to be sub-pixel in scale so smaller dithers are ok.  Knowing this now, I will reduce the dither to 3-pixel or is the program and mount capable of 1-pixel dithers?  As a side note, you would only use larger dithers to eliminate: Hot Pixels, Fixed Pattern Noise, Walking Noise, and Banding.  I only need to dither to help with drizzle of undersampled images.

Update:  Last night I was imaging and I still get spikes when I dither just 3-pixel.  I watched PHD2 as it was dithering ... It looks as if the program is trying to restore the star to its original position and it generates the spikes trying to move back.  The actual move takes only a couple/few seconds.  When I was watching the program during the dither process, it seemed to help to stop guiding and restart so the program selects the new position of the star and stops trying to move the mount back.  When PHD2 received a dither command, shouldn't it stop guiding and then restart after a specified settle time?  

Thanks again Bruce for being patient with all my questions

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/24ed49b1-3437-43dc-ba9f-29dc2a55ae5en%40googlegroups.com.

[Tim Stevenson]

Bruce ... I reviewed the document you sent me concerning dithering.  It states:

"The small shift in pointing position is accomplished  by PHD2 in two steps: 1. Change the lock-point of the primary guide star by the amount requested by the imaging app; and 2. Use the normal guiding routines to move the primary guide star to the new lock-point".  So the program is using spikes to move the mount to the new lock-point and large spikes is due to the large dither?  That makes sense now,  So the key is to use the appropriate settle times to make sure the dither is complete before starting the next image?

[Bruce Waddington]

Hi Tim.  Thanks for taking the time to re-read the User Guide section, I think you understand it correctly now.  There is an option in Advanced Settings for doing "fast re-centers for dithering and calibration" and that option is set by default.  That's what creates the large initial guide pulse after the lock point has been moved.  We think this makes sense as a default setting because it reduces the elapsed time for dithering and settling.  None of the dithering and settling stuff has any effect on the guiding statistics, those things are all ignored for statistical purposes.  So, bottom line, don't make the dithering any larger than you need and just pay attention to the settling behavior.

Cheers,

Bruce

[Brian Valente]

Tim

>>>you would only use larger dithers to eliminate: Hot Pixels, Fixed Pattern Noise, Walking Noise, and Banding.  I only need to dither to help with drizzle of undersampled images.

Dithers address all of these, there isn’t really a strategy of larger vs smaller dithers depending on desired outcomes. 

Brian 

Brian Valente

astronomy portfolio https://www.astrobin.com/users/bvalente/ 

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/3fdf821b-8619-4727-8cfd-c0bf4d72b786n%40googlegroups.com.

[Tim Stevenson]

Hi Brian, I think the dither amount is important and it depends on the size of the noise you are trying to eliminate.  If you had a 15-pixel fixed pattern noise in your sensor, a sub-pixel dither would not eliminate that noise but a sub-pixel dither would not eliminate it.

Bruce, last night I was imaging again with a 3-pixel dither and still had settling failures.  I thought I ran some images with only 1-pixel dithers but I don't see them in my log.  I notice there are a lot of different settle times associated with the mount/PHD2... is there a good explanation of the different settle times?  I assume you could help with any settle times associated with the PHD2 program and I will ask on the NINA Discord forum about their values.

Tim

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/CAJa45i5ADam0MyRBOM0NWPB0Xf0YSc3%3DxhcSk0XL8tfycdCz9w%40mail.gmail.com.

[Brian Valente]

Is your response  hypothetical or actual?

Can you provide an example of a “15 pixel fixed pattern”that would not be resolved by calibration frames?

Brian 

Brian Valente

astronomy portfolio https://www.astrobin.com/users/bvalente/ 

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/CAKrJWDtJmjk2JEN6aLYbovE4zw7vWOGGtmMU2XghF79nQnp6zg%40mail.gmail.com.

[Bruce Waddington]

Tim, please follow the 'Getting Help' instructions and upload both the guide and matching debug log files to our server.  You should do this as a matter of course because you can't know what we will need.  In this case, I can't see what your settling parameters are without that information.

https://openphdguiding.org/getting-help/

Regards,

Bruce

[Tim Stevenson]

https://openphdguiding.org/logs/dl/PHD2_logs_sMQx.zip

here are the log files from last night

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/363703d3-32ea-40b2-a8e6-a3b20f344c93n%40googlegroups.com.

[Bruce Waddington]

Ok, I'll try walk you through one of the typical timeout situations.  You settling parameters here told PHD2 that you wanted star deflections to be less than 1.5px for 10 seconds and you didn't want the entire dither operation to take longer than 40 seconds.  Here's what that looked like:

Point #1 is the start of the dither operation at 03:23:48.  It requests a large dither in both RA and Dec, meaning that the lock-point is "instantly" moved by 9.7 px in RA and 7.4 px in Dec.  This triggers two large initial guide commands that attempt to move the guide star quickly toward that position.  Because you are trying to run with a poor calibration, these initial guide pulses took 8.2 seconds in RA and 2.9 sec in Dec - the poor calibration is why the RA guide pulse is so large.  The initial RA correction resulted in a large over-correction, which then triggered 6 successive RA guide commands to restore order.  The first 5 of these specified a guide pulse of 2.5 sec.  By the time all of this was completed, you had burned through 35 seconds of elapsed time, leaving only 5 seconds left before the timeout limit.  But your settling parameters said that you wanted 10 seconds of guiding with less than 1.5 pixels, so you were simply out of time and triggered the timeout.

Going back to root causes, you got a "poor" calibration result which you didn't do anything about.  The reported guide speed in the mount is 0.9x sidereal which is 13.5 arc-sec/sec.  That part is fine.  But the measured rate of west RA movement seen during calibration was only 6.1 arc-sec/sec, less than 1/2 of what it should have been.  And this is why these large dithers that require large west guide corrections take so long. Smaller dithers that need only east corrections generally settle quickly.   Are you sure the payload is well-balanced in RA on both sides of the pier?  It looks to me like it might not be because the response to west guide commands is poor when the scope is on the east side of the pier.

In order to work through these problems, you should learn to use the LogViewer tool and do as I have done here - zoom in to see the gory details of how the mount is responding during these dither and settling periods.  It can be a time-consuming process so it's not something I can do for you on an ongoing basis.

Good luck,

Bruce