Dithering with Sky Adventurer Mini

[Ian_M]

Hi,

Looking at dithering with my Skywatcher Adventurer Mini (SAM) tracker mount. The latest version of the app has a dithering option that requires an angle (in arcmins) to be entered. Seeking advice/confirmation on how to work out that angle. I am using a Canon 650d (pixel size 4.3um) and a 50mm lens. 

Bit of trigonometry says to me that each pixel subtends angle of ArcTan(4.3/50000) = 0.3 Arcmin. Read somewhere that with a DSLR you should dither 12 pixels which would be approx 4 Arcmin so I am guessing that this is what I should enter into app.

Does that make sense? Any thoughts/commnets greatfully received. Sort of starting from first principles - not sure if they are correct, thoug.

Thanks,

Ian

 

[vlaiv]

What sort of dither movement does application offer? How many subs are you going to capture?

You started off in the right direction - first thing to do is to figure out what is your imaging resolution - or how big pixel is in terms of arc seconds.

Next thing is to figure out movement pattern and size of movement in order not to get to "same place" twice (or at least to reduce probability of that significantly). Third thing of course is that you don't want to crop off too much of your image due to excessive dither movement - remember each time you move a little - you cause edges of that sub to go "out of stack" (or rather intersection between subs shrinks).

Issue with SAM is that it guides in one axis only - RA. This means that any dithers you make will be along one axis which greatly reduces "maneuvering space".

I think that best strategy would be to spend some time to learn how much polar alignment error can you make deliberately before you get trailing in your subs due to poor polar alignment. Here is a good tool to help you with that:

http://celestialwonders.com/tools/driftRateCalc.html

Enter error in polar alignment and observe drift rate. For example - you are imaging at ~18"/px. If your drift rate is 0.1"/second - you can image for 3 minutes before drift makes trail one pixel long.

Why do I mention this? Because if you make polar alignment error such that it trails almost a pixel in single exposure - every exposure will move by one pixel and you will get "dither" in DEC as well. This is sort of natural dither. This makes it almost impossible to get to subs at same location even if RA moves back and forward the same amount (move Ra +5 px, move Dec +1px, move Ra - 5px, move Dec +1px - Ra is at exactly the same place it started but Dec also moved so sub in total moved, but if there was no movement in DEC - subs would align perfectly as +5 -5 = 0).

Btw - you are right - set RA dither around 10-15px (calculated in arc minutes if app requires it like that). I'm guessing here that dither strategy is random and not pattern.

 

Edited April 3, 2020 by vlaiv
typo ...

[Fred_76]

Note that dithering is only available if you have updated the SAM firmware to version 3.10 or above.

I took 10 pixels and arrived to that rule of the thumb:

d' = 35 x p / F

 

Where :

• d is the dithering in arcmin, to enter in the SAM console
• p is the pixel size, in µm
• F is the focal length in mm

You cannot enter decimal numbers directly in the SAM console (at least in iOS). To do that, juste enter 0 and press Done. You will see 0.00 in the box. Then just edit the figures and don't delete  the decimal separator !

I saved a profile for each lens/DSLR combination I use so that I don't need to calculate that again.

Fred

Edited April 7, 2020 by Fred_76

[ursamajorro]

Hi, Fred, it really works using your formula?

[ursamajorro]

I found another formula who seems to be more real and it's better for dithering calculations:

arcsec/pixel = 206,3 x pixel size in um/focal in mm

So, for my Canon dSLR I have for a focal of 135 mm (my old Jupiter lens) I'll have

206,3 x 4,3 / 135 = 6,57 arcsec/pixel

and if I want to dither 10-15 pixels I should set my dithering field in SAM console to 6,57 x 10 or 15 = from 65,7 until 98,55 arcseconds, ie 1,1 or 1,64 arcmin.

[spordniar]

On 29/04/2020 at 05:41, ursamajorro said:

I found another formula who seems to be more real and it's better for dithering calculations:

arcsec/pixel = 206,3 x pixel size in um/focal in mm

So, for my Canon dSLR I have for a focal of 135 mm (my old Jupiter lens) I'll have

206,3 x 4,3 / 135 = 6,57 arcsec/pixel

and if I want to dither 10-15 pixels I should set my dithering field in SAM console to 6,57 x 10 or 15 = from 65,7 until 98,55 arcseconds, ie 1,1 or 1,64 arcmin.

I think Fred's formula is exactly the same as yours, just simplified.

(206.3 * 10) * pixel size / focal len = arcsec
(206.3 * 10 / 60) * pixel size / focal len = arcmin
34.383  * pixel size / focal len = arcmin

So I guess 35 give you approximately  10 pixel dither, and if you want 15 pixel just multiply by 1.5.

Nice post guys!

[Fred_76]

@spordniar is absolutely right ! The formula are the same, the one I give is just simplified 😉

However I found that using 30 pixels instead of 10-15 gives better results. Therefore the formula becomes :

 

d' = 100 x p / F

Where :

• d is the dithering in arcmin, to enter in the SAM console
• p is the pixel size, in µm
• F is the focal length in mm

I do not agree on the statement of @vlaiv when he says :

Quote

if you make polar alignment error such that it trails almost a pixel in single exposure - every exposure will move by one pixel and you will get "dither" in DEC as well. This is sort of natural dither.

As the drift in DEC will always be in the same direction and at the same rate, it will generate a trail, responsible of walking noise pattern. It will not generate any "natural dither". To reduce the countereffect of the drift in DEC, when dithering on RA only, you have to dither a lot, so the 30 px instead of the 10-15 usual pixels that are commonly used when dithering in both directions.

Note that the standby time between two photos will be affected by the dithering operation. As the Star Adventurer Mini can only move at x0-x1-x2 the sidereal speed, the duration in seconds of a dithering jump will be at max four times the arc.min set in the app. For exemple if you set 5 arc.min, the duration of a dithering will be 4x5=20 s (this is a max value as each dithering jump will be random between 0 and 20 s) + some seconds to stabilize, before the next photo starts.

[vlaiv]

8 minutes ago, Fred_76 said:

As the drift in DEC will always be in the same direction and at the same rate, it will generate a trail, responsible of walking noise pattern.

 

On 03/04/2020 at 16:56, vlaiv said:

Why do I mention this? Because if you make polar alignment error such that it trails almost a pixel in single exposure - every exposure will move by one pixel and you will get "dither" in DEC as well. This is sort of natural dither. This makes it almost impossible to get to subs at same location even if RA moves back and forward the same amount (move Ra +5 px, move Dec +1px, move Ra - 5px, move Dec +1px - Ra is at exactly the same place it started but Dec also moved so sub in total moved, but if there was no movement in DEC - subs would align perfectly as +5 -5 = 0).