Contemplating starting a telescope design software project. Thoughts?

[pipnina]

I've just finished the second year of a foundation degree in software engineering. I've been thinking of doing a project over the summer, but one that would have a practical use (i.e. a piece of software that people would want to use when it is "done").

I was thinking about creating a program for designing telescopes: It is complicated but can have simple beginnings, some amateurs could find it useful for DIY projects or just to test ideas. I would be starting with mirrors first, as I would not have to consider the multitude of refractive indexes given by the many types of glass.

I have found a few programs freely available on the internet that could accommodate the DIY amateur, but so far as I could see they were all quite old and poorly maintained, or with restrictions (i.e. limited to mirrors only, refracting only, component limit etc) or even just not great pieces of software to begin with.

What do you guys think, worth a go? Would you guys find much use or enjoyment out of such a program?

[wornish]

Not sure how many people could make use of such a program to be honest. your target user base would be very small.  

How many people can make their own mirrors or lenses its a very specialised and expensive process isn't it?

Sorry to pour cold water on your idea others might disagree, or perhaps I have misunderstood the goal of the software.

 

[pipnina]

47 minutes ago, wornish said:

Not sure how many people could make use of such a program to be honest. your target user base would be very small.  

How many people can make their own mirrors or lenses its a very specialised and expensive process isn't it?

Sorry to pour cold water on your idea others might disagree, or perhaps I have misunderstood the goal of the software.

That's fair criticism, it is rather specialized. If you've other ideas though (maybe software you've pined for personally) I'd be happy to consider it.

[Gasman]

Made a couple of mirrors myself in the past, hugely enjoyable activity but not sure whereabouts you could acquire materials for it nowadays?. The software could be useful if the materials are easily available. Telescope control and image processing seem to be the things at present!.

Steve

[Sunshine]

i have been contemplating designing my own Dob for some time, i will start this winter as i am off 2 month in winter, i think everyone could benefit from software fir that purpose!

[Dr_Ju_ju]

The one bug-bear of every imager, Polar Alignment.

I'm not talking about software for manual adjustment as already exists, but a system with stepper motors that can adjust both Alt\Dec axes, i.e. an automated 'Polemaster' with stepper drivers etc.... 

[pipnina]

17 minutes ago, Dr_Ju_ju said:

The one bug-bear of every imager, Polar Alignment.

I'm not talking about software for manual adjustment as already exists, but a system with stepper motors that can adjust both Alt\Dec axes, i.e. an automated 'Polemaster' with stepper drivers etc.... 

I admit I would also find such a thing useful, however I would not have the capability to develop the hardware for it. Even if I could, I would need to write drivers to go alongside the control software.

[Stub Mandrel]

I would like one that simplified making a refractor using a pre-existing lens, calculating thing like how long the drawtube should be, how much it should move and how far from the lens it should be.

You could include information on vignetting for different size sensors/eyepieces.

(Currently 3D printing an adaptor to joining an old cheap plastic synta focus mechanism to an aluminium tube, to be followed by a holder for a 50 x 288mm lens to make my third home brew guidescope!)

[cgarry]

23 hours ago, pipnina said:

I've just finished the second year of a foundation degree in software engineering. I've been thinking of doing a project over the summer, but one that would have a practical use (i.e. a piece of software that people would want to use when it is "done").

In my experience you do not just write software and it is "done"!  You either continuing maintaining it forever or it will be considered "no longer maintained" and it will die a natural death...

Cheers,
Chris

[pipnina]

Just now, cgarry said:

In my experience you do not just write software and it is "done"!  You either continuing maintaining it forever or it will be considered "no longer maintained" and it will die a natural death...

Cheers,
Chris

This is true, by "done" (hence quotes) I meant "feature complete" (i.e. has the features planned). If the software is still used, and I have the time (assuming it gets to a "feature complete" state in the first place of course) I don't see why it couldn't be maintained in either my spare time or by any other devloper (I would be releasing it under GPL3, so other people could take the source code, modify it, redistribute it and whatever.).

Ultimately with software projects of that nature, if they are desired, they will survive. If they are unwanted, they will fade into obscurity and probably stop working in a few years as computers and operating systems move on.

[cgarry]

If you are planning a career in software engineering, or even a similar field, then having an open source project such as this under your belt actually makes you very employable.  The code itself showcases your skills and the fact that you have done it shows that you have drive. 

Cheers,
Chris

 

[vlaiv]

Depending what type of software are you going to build.

I've found that geometrical optics software is straight forward. Not difficult at all. Physical optics on the other hand requires quite a bit of knowledge both in physics/math and advanced programming.

I was thinking of doing some simple software for calculation of PSFs for example, and although it can be done with Fourier transforms of aperture - things like off axis / different mirror shapes gets really complicated really fast.

Anyway I think it is a good idea, and if you are going to work on it as open source, I might join in (if you are OK with that).

 

[pipnina]

23 hours ago, vlaiv said:

Depending what type of software are you going to build.

I've found that geometrical optics software is straight forward. Not difficult at all. Physical optics on the other hand requires quite a bit of knowledge both in physics/math and advanced programming.

I was thinking of doing some simple software for calculation of PSFs for example, and although it can be done with Fourier transforms of aperture - things like off axis / different mirror shapes gets really complicated really fast.

Anyway I think it is a good idea, and if you are going to work on it as open source, I might join in (if you are OK with that).

I had envisioned the software providing on-axis and off-axis star images. I had presumed it might require a ray-tracing approach though I had no idea how diffraction would be accounted for. 

Collaboration sounds ideal! I have not begun any work towards it yet, but once I've drafted up a more precise idea of what I envisioned the software being (sketches, wireframes etc) I could message you about it? Then maybe we can bounce ideas around.

I don't know exactly what kind of programming you are familiar with, but I had planned to create the software using C++ and the QT toolkit, and graphics stuff (i.e. drawing the OTA, rendering the star images) being done using OpenGL. This is mostly out of the interest of platform cross-compatibility. DirectX would limit the program to Windows, and Vulkan is a tad too complicated.

[vlaiv]

16 minutes ago, pipnina said:

I had envisioned the software providing on-axis and off-axis star images. I had presumed it might require a ray-tracing approach though I had no idea how diffraction would be accounted for. 

Collaboration sounds ideal! I have not begun any work towards it yet, but once I've drafted up a more precise idea of what I envisioned the software being (sketches, wireframes etc) I could message you about it? Then maybe we can bounce ideas around.

I don't know exactly what kind of programming you are familiar with, but I had planned to create the software using C++ and the QT toolkit, and graphics stuff (i.e. drawing the OTA, rendering the star images) being done using OpenGL. This is mostly out of the interest of platform cross-compatibility. DirectX would limit the program to Windows, and Vulkan is a tad too complicated.

Yes, sure, PM me about discussion anytime.

C++ and QT are fine (although I haven't used QT in some time, but I also started a few project in this exact setup, so I'm bringing my self up to speed with it). Don't think you need to go OpenGL route (again fine with me I did a lot of 3D stuff back in the day, both in DirectX and OpenGL), but that you should decide based on which type of modeling you are going to do - most if not all things related to telescope optics can be presented in 2D due to symmetrical nature of instruments - so simple canvas drawing will suffice.

If you want to do star images on/off axis, defocused stars and such as they would appear in image, not just spot diagrams, then we are talking physical optics. I'll dig up some nice papers that explain how things work in that approximation, but for algorithms, I'm afraid we are on our own (you should not be worried though, I've got 20+ years experience as software engineer  / system architect so I believe we can manage it).

In the mean time take a look at this for example, just to briefly introduce yourself to physical optics (if you are not familiar with subject yet):

http://www.photonics.intec.ugent.be/download/ocs130.pdf

 

[Guest chaz2b]

I`m with Neil, a more practical direction for those without the genious of Sheldon!, the use of 3D printers is the latest vibe. I like to make things ,but the practicalities of actually making the optics is outside of most peoples pockets imho.

To reiterate what Neil has said, figuring out tube lengths/ratios/draw tubes etc would be more appealling.

 

chaz

[GreatAttractor]

I've played with raytracing (i.e. geometrical/ray optics only) a bit, indeed it's straightforward. Building a full-fledged optics simulation tool with nice 3D GUI would take a lot of time, though. Still, a simple command-line prog accepting a textual description of optics is enough for some simple experiments (e.g. define point light sources spaced on a grid, “image” them through an overly short-f.r. Cassegrain and make a time lapse out of it around the focal plane, illustrating the distortion and field curvature).

Ping me if you decide to go down this route and have trouble with deriving formulas for reflection off paraboloid or hyperboloid... ?‍?