Yet another DIY Electronic collimator

[rsarwar]

I am planning to get a 200 mm RC scope but i hate collimation process on my newtonian. i tried lasers and a cheshire. whilst i am able to get fairly okay collimation following testing with a LED torch under a Al foil with a pin hole, the process usually takes hours. and the results were never fully satisfactory . then i saw a Ocal Electronic Collimator on FLO and thought it should not be difficult to make it using a Arducam + some python/openCV coding. So I 3d  printed a 2 inch frame in which I screwed the camera. 

first impression, need to  reprint the 2 inch frame in red, so that it is easier to see on the mirror. camera i chose is a bit expensive but works well.

total cost was about 50 pounds for the camera. python codes and stl files are in https://github.com/rsarwar87/pyReflectorCollimator

camera: https://www.amazon.co.uk/gp/product/B07YHJK4LN/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1

Edited November 28, 2021 by rsarwar

[JamesF]

I have to admit that I wasn't aware of DIY electronic collimators at all, so "yet another" came as a surprise.  Interesting project though.

James

[rsarwar]

 

4 hours ago, JamesF said:

I have to admit that I wasn't aware of DIY electronic collimators at all, so "yet another" came as a surprise.  Interesting project though.

James

I am sure someone must have made it. 

Edited November 27, 2021 by rsarwar

[markse68]

I’m trying to work out what the Ocal is- is it just a camera with concentric circle overlay or does it do something cleverer? I have used my PA camera- an ASI120, with its wide angle lens and freeware concentric circles overlay which is quite useful for adjusting secondary position.

Mark

[kbrown]

Good job! I've done something similar with my QHY5L-II and a CS mount CCTV lens attached to it. I used oacapture to view the video feed. It has a few different style reticles you can overlay on top of the image but nothing like in your software. Curious if you're planning to add support for any commercial astro cameras?

[johninderby]

42 minutes ago, markse68 said:

I’m trying to work out what the Ocal is- is it just a camera with concentric circle overlay or does it do something cleverer? I have used my PA camera- an ASI120, with its wide angle lens and freeware concentric circles overlay which is quite useful for adjusting secondary position.

Mark

It is a lot more complex than that. 

http://www.ocalworld.com/en/PDF说明书/OCAL Electronic Collimator PRO user's Manual.pdf

Still expensive though. 🤔

[rsarwar]

14 hours ago, kbrown said:

Good job! I've done something similar with my QHY5L-II and a CS mount CCTV lens attached to it. I used oacapture to view the video feed. It has a few different style reticles you can overlay on top of the image but nothing like in your software. Curious if you're planning to add support for any commercial astro cameras?

I could. Should not be difficult at all. However, how do you change focus on the Cs mounted lens? 

I have a ASI290mm with the c mount lens it comes with, but the reason why I got the camera for arducam is because we need to be able to change focus. This would allow you to correctly focus properly on the edges of the second art mirror to ensure that it is within the circular overlay. And then swap focus on to the image on the secondary mirror to fine tune the secondary and align the primary mirror. 

 

Ps: zwo publishes python drivers for their camera, so it will be straight forward indeed

https://github.com/python-zwoasi/python-zwoasi/blob/master/zwoasi/examples/zwoasi_demo.py

 

Edited November 28, 2021 by rsarwar

[rpineau]

@rsarwar , which Arducam are you using exactly (they have a few ).

Do you have a link ?

Thanks

Rodolphe

 

[rsarwar]

9 minutes ago, rpineau said:

@rsarwar , which Arducam are you using exactly (they have a few ).

Do you have a link ?

Thanks

Rodolphe

 

This one: https://www.amazon.co.uk/gp/product/B07YHJK4LN/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1

but i suppose you can also get the ones that are based MIPI, but that may require modifying the application.

basically any camera that can have motorized focus would do and can be centred on the two inch would work.

Edited November 28, 2021 by rsarwar

[kbrown]

6 hours ago, rsarwar said:

I could. Should not be difficult at all. However, how do you change focus on the Cs mounted lens?

I didn't really bother doing much with the focusing. I just set it manually to where I thought I'd see enough. I used this mainly for centring and aligning the secondary mirror on my newt after taking it out. Really handy for that. Then I used a laser collimator and a chesire eyepiece to fine tune things. Iterated the whole process a few times and I got a very good collimation.

[rsarwar]

46 minutes ago, kbrown said:

I didn't really bother doing much with the focusing. I just set it manually to where I thought I'd see enough. I used this mainly for centring and aligning the secondary mirror on my newt after taking it out. Really handy for that. Then I used a laser collimator and a chesire eyepiece to fine tune things. Iterated the whole process a few times and I got a very good collimation.

I see. I guess that is the most tricky bit, however, If you are able to change the focus you should be able to do the whole collimation process without a laser. Like collamating with a Cheshire only. Saves you from ensuring correct collimation the laser collimator as well.

 

I just used the laser to confirm that collimation worked. 

Edited November 28, 2021 by rsarwar

[malc-c]

Interesting.  I remembered Dion using Sharpcap and an overlay application to do something similar about a decade ago

 

 

[Mattynorth]

On 29/11/2021 at 08:15, malc-c said:

Interesting.  I remembered Dion using Sharpcap and an overlay application to do something similar about a decade ago

 

 

A decade ago! That made me realise I’ve been doing this for quite a while now….and I still don’t know what I’m doing lol.

[rsarwar]

I think i managed to work out a "proper" way of collimating a RC. The idea is to 

1. make the secondary and the primary coaxial without the focuser attached.

2. add M90 tilt plate and the focuser and then by only adjust the collimation on the tilt plate to bring the focuser and the secondary coaxial. 

first the first one, i had to 3d print a M90 to M48 thread.

 

The adator was made such that it uses the pressure lock ring to centre the adaptor. and the on the M48 thread, screw in the electric collimator. then reiterate the collimation on the secondary and primary mirrors to get it as close as possible.

then take it off, put the tilt ring and the focuser and then the electric collimator and then adjust the tilt plate to recollimate. 

I dont see a point of perfectly dialing in the screws, but as close to perfect is good enough, as this will need to be readjusted using a artifitial star. I dont have one so i did it under the stars, but forgot to take images to document it. however, the above process resulted in near perfect collimation and needed on mirror adjustment.

Edited January 30, 2022 by rsarwar

[Leeps]

Just a check-in here to say that this is extremely cool. I'm hoping to try it out and contribute, but the "HD Autofocus webcam" I ordered is 480p, and manual focus 🤣It also died immediately when I turned the manual focus screw and it backed the lens into the sensor. Once I have a webcam I'll try this out

[zernikepolynomial]

I cloned your project and decided to make a few adjustments to make it look like the first version. I think the latest version was not rendering correctly. Anyway, I am going to print out a PLA prototype at my local shop. 

What kind of screws for the standoffs did you use? And print method/material did you use? Did you use self tapping screws and FDM print method?

 

I was thinking maybe heat inserts could be put in the bottom holes. I think the top holes would be too delicate and it would ruin the standoffs.

[kamayok3]

On 26/11/2021 at 09:13, rsarwar said:

I am planning to get a 200 mm RC scope but i hate collimation process on my newtonian. i tried lasers and a cheshire. whilst i am able to get fairly okay collimation following testing with a LED torch under a Al foil with a pin hole, the process usually takes hours. and the results were never fully satisfactory . then i saw a Ocal Electronic Collimator on FLO and thought it should not be difficult to make it using a Arducam + some python/openCV coding. So I 3d  printed a 2 inch frame in which I screwed the camera. 

first impression, need to  reprint the 2 inch frame in red, so that it is easier to see on the mirror. camera i chose is a bit expensive but works well.

total cost was about 50 pounds for the camera. python codes and stl files are in https://github.com/rsarwar87/pyReflectorCollimator

camera: https://www.amazon.co.uk/gp/product/B07YHJK4LN/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1

I am having an issue running your .py files in Windows 10/11. I already have Python 3.11 installed.  Any guide on how to run it?

 

Best Regards,

Tom

[rsarwar]

Hi it was written to work from linux, but will work on rpi4. May also work on wsl.

[kamayok3]

On 19/08/2023 at 05:40, rsarwar said:

Hi it was written to work from linux, but will work on rpi4. May also work on wsl.

Is it possible to change video resolution inside your apps?

[rsarwar]

On 21/08/2023 at 15:00, kamayok3 said:

Is it possible to change video resolution inside your apps?

Should be fairly straight forward as it is using opencv. Here is how

https://www.google.com/amp/s/www.tutorialkart.com/opencv/python/opencv-python-resize-image.amp/

 

1.Multiply with scaler to resize resolution

https://github.com/rsarwar87/pyReflectorCollimator/blob/master/pyCameraController.py#L63-L69

2. Resize frame by the same scalar after this line

https://github.com/rsarwar87/pyReflectorCollimator/blob/master/pyCameraWindow.py#L53

        ret, frame = self.cc.vid.read()

[rjsc2001]

Nice project.

What webcam can be used? Must be able to focus on windows and cheap.

Most webcams don't say if they can be focused in windows or not...

[Spile]

If this could be run an an esp32 cam it would make a cost effective wireless project.

[rsarwar]

you can use any camera that works on linux/rpi and can be opened using opencv in python. all that needs doing is changing the "video_idx" here  so that opencv can open it.

 

the reason for choosing the camera i chose is because having the ability to change focus is critical to ensure smooth calibration without needing to touch the camera and move it.