Dithering with PHD2 - setteling issues

Cameras, telescopes, focusers and other hardware
Post Reply
User avatar
Rudi
Posts: 152
Joined: 08 Jan 2019, 04:47

Dithering with PHD2 - setteling issues

Post by Rudi » 15 Mar 2020, 08:30

I tried the dithering feature a couple of month back, and that night, it worked fine, but it was only for a few frames.

Last night, we had our first clear nights for months here in Denmark, so I thought, that I would use dithering for the whole session.

My setup:
Mount: NEQ-6 Pro
Telescope: Sky-Watcher Quattro 8" f/3.9 + Sky-Watcher Aplanatic Corrector
Camera: ASI 1600 MM Pro + ASI 8 pos FW + EAF
Guidescope: 50 mm f/5
Guidecam: ASI120 MC
Guide SW: PHD2 v.2.6.5
AstroArt 7 SP2

L,R,G,B exposure times 120 sec
PHD2 exposure time 2 sec
AstroArt dithering 6x6 pixels after each image

I have ensured to select "Send message to PHD Guiding" in AA7.

The issue is, that I had to ditch 33 of 44 images !!!
It seems, that PHD2 had a tough time setteling properly.

Is there anything I can do about this?

Notice the effect the "setteling issue" has on images, teh image shifts:
Image

PHD2 log extract from when the setteling is ok:
Image

PHD2 log extract from when the setteling is bad:
Image
/Rudi

fabdev
Posts: 461
Joined: 03 Dec 2018, 21:43

Re: Dithering with PHD2 - setteling issues

Post by fabdev » 15 Mar 2020, 11:58

Dithering with " 6 " is very high, it looks like the settling time was more than 15 seconds. Several pulses in DEC were not sufficient to take the star back, and the guide in DEC seems too "lazy", also during normal guide.

Maybe last time the amount of dithering was lower? Do you really need 144 dithering positions? (12x12, if in PHD "6" is the radius, I'm not sure). If yes, you can solve using Spiral dithering which moves the dithering position only by 1 pixel at a time, this will solve and (at least in Astroart) it's the best algorithm. The culprit is also Astroart, where the maximum settling time expected from PHD is 5 seconds, then the exposure is started anyway.

About the fixed dithering: in Astroart dithering with "6" means a square of 12x12= 144 positions, which are visited randomly but with a maximum distance of 8-9 pixels. In PHD, "6" could have a different meaning though. In Astroart "Spiral" is always better, but remember that this option must be set inside PHD, not Astroart, since it's not transmitted (when "Send message to PHD guiding" is enabled, the option is grayed).

User avatar
Rudi
Posts: 152
Joined: 08 Jan 2019, 04:47

Re: Dithering with PHD2 - setteling issues

Post by Rudi » 15 Mar 2020, 12:41

Thanks for the answer Fabio - I increased the dithering dx/dy to 6, because I knew that I needed arround 50 frames. But is it ok to have less dithering positions than the number of frames?

And yes, my declination has quite some backlash, but I hoped, that AstroArt and PHD could "talk together" and wait until the declination had settled. i will see, if I can change that in PHD2 itself?
/Rudi

fabdev
Posts: 461
Joined: 03 Dec 2018, 21:43

Re: Dithering with PHD2 - setteling issues

Post by fabdev » 15 Mar 2020, 13:11

It depends on the camera, but I think that if you lose 15 seconds every 120 seconds (so 12% of "quantum efficency" is lost) you are ruining the quality of the final image instead of improving with dithering (e.g. in the same time you could have taken more images).

Again, it depends on the camera and the quality of the dark frame, but I think that a dithering of "1" (2x2, 4 positions) and "2" (4x4, 16 positions) would give the same result on most cameras.
For a sequence of 50 images I would use "2", so 16 positions, 1 shift every 4 images, with spiral.
As a general rule, I would not lose more that 2-3% of exposure time for doing dithering, on most cameras this will break even. Actually dithering was used at the beginning for taking images without dark frames, then for improving photometry.

I made some mistakes in the last message. In Astroart the unit is a radius, so "6" gives 12x12 = 144 dithering position. I will correct it. Yes a better communication from Astroart would have prevented the problem. I also confirm that all settings (radius value and spiral mode, MUST be set inside PHD).

Post Reply