Adding a surface to an existing model

[joelmartis]

I am trying to combine a few optical lenses from Thorlabs, Edmunds etc. I don't see an easy way to simply combine the surfaces in two models (i.e. the thorlabs lens followed by the edmunds lens), so I loaded the first model and decided to manually add the surfaces from the other model. However this turns out to not work:

My code:

#%% I import Thorlabs part https://www.thorlabs.com/thorproduct.cfm?partnumber=PAF2-A4A

from rayoptics.environment import *

root_pth = Path(rayoptics.__file__).resolve().parent

opm, info = open_model("...relevant_directory/TTN136376.ZMX", info=True)

sm  = opm['seq_model']
osp = opm['optical_spec']
pm = `opm['parax_model']`
em = opm['ele_model']
pt = opm['part_tree']
ar = opm['analysis_results']


osp['pupil'] = PupilSpec(osp, key=['object', 'pupil'], value=1)
osp['fov'] = FieldSpec(osp, key=['object', 'height'], value=0, flds=[0]) #point source

sm.gaps[0].thi=3

opm.flip() #flipping the lenses


opm.update_model()


layout_plt = plt.figure(FigureClass=InteractiveLayout, do_draw_rays=True, do_paraxial_layout=False, opt_model=opm)
layout_plt.refresh()

#so far so good!

#%% Now I try to manually add the second lens from Edmunds

sm.add_surface([0.064499, 1.6, 'N-BK7', 'Schott'])
sm.add_surface([0.0, 30.0])

sm.list_model() #the output doesn't make sense, the surfaces haven't been added correctly

Output:
c t medium mode zdr sd
Obj: 0.000000 3.00000 air 1 0.0000
1: 0.127065 1.30000 N-LASF9 1 1.0000
2: 0.458716 1.11000 N-SK16 1 1.0000
Stop: -0.420168 5.00000 air 1 1.0000
4: -0.000000 2.80100 air 1 0.90000
Img: 0.000000 1.60000 N-BK7 1 0.32818
6: 0.064499 30.0000 air 1 1.0000
7: 0.000000 0.00000 1 1.0000

---

Your help in correcting this is much appreciated!

Also, once a surface has been added, is there a way to quickly edit it? Such as changing the curvature, thickness or material? And how about being able to add a surface between two existing surfaces?

Thanks,
Joel

[mjhoptics]

Hello @joelmartis,
Thanks for your question. I've worked an example below; things to highlight include setting the current surface for the add_surface() command, a workaround to a ray-optics bug that InteractiveLayout is missing elements and how to modify the system after it's been entered.
There is also an example in the doc called Multiple Lens Import Example
Please feel free to follow up with any questions.
Regards
Mike Hayford

Investigate issue #141

from rayoptics.environment import *

opm, info = open_model(Path().parent / "PAF2-A4A-Zemax/TTN136376.ZMX", info=True)

sm  = opm['seq_model']
osp = opm['optical_spec']
pm = opm['parax_model']
em = opm['ele_model']
pt = opm['part_tree']
ar = opm['analysis_results']

osp['pupil'] = PupilSpec(osp, key=['object', 'pupil'], value=1)
osp['fov'] = FieldSpec(osp, key=['object', 'height'], value=0, flds=[0]) #point source

sm.gaps[0].thi=3

opm.flip() #flipping the lenses

opm.update_model()

layout_plt = plt.figure(FigureClass=InteractiveLayout, do_draw_rays=True, do_paraxial_layout=False, opt_model=opm).plot()

#so far so good!

We need to know the insertion point in the sequence that add_surface will use.

sm.cur_surface

5

sm.list_model()

              c            t        medium     mode   zdr      sd
  Obj:     0.000000      3.00000       air             1      0.0000
    1:     0.127065      1.30000   N-LASF9             1      1.0000
    2:     0.458716      1.11000    N-SK16             1      1.0000
 Stop:    -0.420168      5.00000       air             1      1.0000
    4:    -0.000000      2.80100       air             1     0.90000
  Img:     0.000000      0.00000                       1     0.32818

Surface 5 is the image surface. We want to insert after, say, the dummy plane - surface 4.

sm.set_cur_surface(4)

While adding the surfaces, set circular clear apertures to match the Thorlabs OD

sm.add_surface([0.064499, 1.6, 'N-BK7', 'Schott'])
sm.ifcs[sm.cur_surface].clear_apertures.append(srf.Circular(1))

sm.add_surface([0.0, 30.0])
sm.ifcs[sm.cur_surface].clear_apertures.append(srf.Circular(1))

sm.list_model()

              c            t        medium     mode   zdr      sd
  Obj:     0.000000      3.00000       air             1      0.0000
    1:     0.127065      1.30000   N-LASF9             1      1.0000
    2:     0.458716      1.11000    N-SK16             1      1.0000
 Stop:    -0.420168      5.00000       air             1      1.0000
    4:    -0.000000      2.80100       air             1     0.90000
    5:     0.064499      1.60000     N-BK7             1      1.0000
    6:     0.000000      30.0000       air             1      1.0000
  Img:     0.000000      0.00000                       1     0.32818

opm.update_model(src_model=sm)

As identified in several other issues, a defect in ray-optics fails to recognize that adding surfaces should add lens elements in the new layout. You have to force ray-optics to rebuild the list of elements it uses for rendering by calling the opm method rebuild_from_seq().

opm.rebuild_from_seq()

layout_plt1 = plt.figure(FigureClass=InteractiveLayout, do_draw_rays=True, do_paraxial_layout=False, opt_model=opm).plot()

create a cemented doublet by inserting a surface between 5 and 6

Insertion using add_surface() inserts the new surface after the cur_surface. This is different from Python list insertion but follows the practice of most image forming software.

sm.set_cur_surface(5)
sm.add_surface([-0.64499,0.9,'N-SF6', 'Schott'], sd=1)

opm.update_model()
opm.rebuild_from_seq()

layout_plt2 = plt.figure(FigureClass=InteractiveLayout, do_draw_rays=True, do_paraxial_layout=False, opt_model=opm).plot()

sm.list_model()

              c            t        medium     mode   zdr      sd
  Obj:     0.000000      3.00000       air             1      0.0000
    1:     0.127065      1.30000   N-LASF9             1      1.0000
    2:     0.458716      1.11000    N-SK16             1      1.0000
 Stop:    -0.420168      5.00000       air             1      1.0000
    4:    -0.000000      2.80100       air             1     0.90000
    5:     0.064499      1.60000     N-BK7             1      1.0000
    6:    -0.644990     0.900000     N-SF6             1     0.39164
    7:     0.000000      30.0000       air             1      1.0000
  Img:     0.000000      0.00000                       1     0.32818

Editing the sequential model

The individual attributes of the sm can be changed directly. It must be followed up by calling update_model() on the opm.

The Sequential Model is organized around lists of interfaces and gaps. Interfaces are commonly implemented as Surfaces; Surfaces support profile definitions and interface with the ray trace. Gaps consist of a thickness (the separation between adjacent interfaces) and an optical material specification.

sm.gaps[5].medium = create_glass('N-SK16', 'Schott')
sm.gaps[5].thi = 1.2
sm.ifcs[7].profile.cv = -.12

opm.update_model()
opm.rebuild_from_seq()

sm.list_model()

              c            t        medium     mode   zdr      sd
  Obj:     0.000000      3.00000       air             1      0.0000
    1:     0.127065      1.30000   N-LASF9             1      1.0000
    2:     0.458716      1.11000    N-SK16             1      1.0000
 Stop:    -0.420168      5.00000       air             1      1.0000
    4:    -0.000000      2.80100       air             1     0.90000
    5:     0.064499      1.20000    N-SK16             1      1.0000
    6:    -0.644990     0.900000     N-SF6             1     0.39570
    7:    -0.120000      30.0000       air             1      1.0000
  Img:     0.000000      0.00000                       1     0.32818

layout_plt3 = plt.figure(FigureClass=InteractiveLayout, do_draw_rays=True, do_paraxial_layout=False, opt_model=opm).plot()

[joelmartis]

Thank you Michael! This is very helpful!

I had another, not related to this particular issue.

Can I export a ray trace as a wavefront? I want to do some diffraction limited spot size simulations, and exporting the ray trace as a wavefront (i.e. angle v/s phase or something equivalent) would be very helpful.

Joel

[mjhoptics]

Hi Joel,
I have a couple of jupyter notebooks that demonstrate wavefront calculations. They are in my notebook repo.

https://github.com/mjhoptics/ray-optics-notebooks/focus%20tilt%20notebook.ipynb
https://github.com/mjhoptics/ray-optics-notebooks/wavefront%20notebook.ipynb

Both of these notebooks setup interactive widgets to change defocus and image position to demonstrate their effect on the wavefront.
Hope these help.
Mike

[joelmartis]

Thank you so much Michael! This is very helpful

…

-Joel

On Tue, Mar 12, 2024 at 9:21 AM Michael Hayford ***@***.***> wrote: Hi Joel, I have a couple of jupyter notebooks that demonstrate wavefront calculations. They are in my notebook repo. https://github.com/mjhoptics/ray-optics-notebooks/focus%20tilt%20notebook.ipynb https://github.com/mjhoptics/ray-optics-notebooks/wavefront%20notebook.ipynb Both of these notebooks setup interactive widgets to change defocus and image position to demonstrate their effect on the wavefront. Hope these help. Mike — Reply to this email directly, view it on GitHub <#141 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACPSR4DGGZJOX7EKJDTNZATYX4TRXAVCNFSM6AAAAABEOK5VM2VHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DONRSGMYDE> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[joelfrancis1122]

yeah it is

[ebflagg]

This was very helpful because the existing documentation does not describe how to modify models after they have been created. However, when I followed the example above, I found that the layout plot was not rendering correctly until I moved the rebuild_from_seq() method to run before the update_model() method. I'm using rayoptics==0.9.2, which is the current version as of this date (11/15/24).

[mjhoptics]

Hello @ebflagg,
I think the script above will work without any calls to rebuild_from_seq(). Version 0.9 had major revisions to keep the sequential model and element model in sync. Everything should be handled by the update_model() call.
Hope this helps.
Mike Hayford

[ebflagg]

@mjhoptics
You are correct! I had included rebuild_from_seq() because it was in the example above. But now I've removed it entirely and the layout plots are correct just from running update_model() after modifying the sequential model. In fact, if I run rebuild_from_seq() without then running update_model() the layout plot becomes incorrect again. (Later-added lenses appear overlapping earlier-added lenses.) So, something might still be wrong with rebuild_from_seq(). But since update_model() takes care of the updating, we don't really need rebuild_from_seq()---at least in the notebook I wrote to practice modifying lenses.