34 replies to this topic

[patta]

I've assembled one "Flint glass" spectrograph,  starting from the recipe as from post in the "Optical designs" thread:  https://www.cloudyni...ead/?p=13954532

Here the lenses and M42 adapters

.

.

The first trial gave a lot of coma and astigmatism; likely because the cemented triplets are different than expected. rearranged the optics as follow, with an extra singlet lens. new arrangement, less coma, but also less chromatic dispersion

.

.

Testing with APS-C camera; and a 50mm objective; test target RGB led strips

.

Edited by patta, 04 March 2025 - 02:14 PM.

[patta]

Results on test bench:

 

Without slit, imaging small white LEDS:

.

.

With the 9.5mm slit, imaging a Neon tube lamp:

.

.

 

Not much more color separation than the original at post #1 above; but has larger aperture etc.

Edited by patta, 04 March 2025 - 01:55 PM.

[patta]

Here some rough analysis of the performance:

 

With the 9.5mm diameter, 25 micron wide slit, 

the relay work at about 1.5 x magnification

Aperture 3 mm, about f/32 on image

Projecting on 18 MPx APS-C sensor 

looking at the Neon lamp:

.

.

And here the spatial resolution - the Neon lamp has a convenient dot pattern!

.

.

A single red LED (no slit)

.

 

Spatial (tangential) resolution, is good, I'm quite happy

 

Spectral resolution is a bity underwhelming - but maybe in the UV...

In the first image, I assume the last peak in violet is in best focus, and what we see is the width of the slit. Maybe a thinner slit will improve the resolution?

 

The sensor (Canon 600D) has pixel size 4.3 micron

At f/32, diffraction- limit for 550 nm wavelength is ~15 micron

Ballpark of resolution from the above images:

Spatial 15 - 25 pixel -> 60 - 100 micron

Chromatic resolution: 20-40 pixel -> 70-120 micron

Slit width projected 25x1.5 = 35 micron

 

Cromatic dispersion 450-650 nm: 400 px  -> resolves 10-20 bands

Circumference of slit image: 6000 pixels   -> resolves 500-1000 points

 

 

 

 

Resolution is way too low. why??

Edited by patta, 04 March 2025 - 02:51 PM.

[patta]

Here check if the system transmit UV

 

Light from LED with peak at 390 nanometers pass es no problem.

Light from UV LEDs emitting 365 nm does not pass - only the fluorescence at longer wavelengths is visible.

Maybe is the optics, maybe is the sensor filter. 360 nm is the hard limit where flint glass does not transmit anymore.

 

Anyway, the dispersion in the UV around 390 nm (Calcium K...) is substantial, so some spectrometry could be done there. 

Now I need some spectral lines in the UV....

.

Edited by patta, 04 March 2025 - 02:21 PM.

[patta]

Took a few more photos of the Neon lamp; maybe I've messed up with the aperture calcuation, was too narrow? Yes, I've forgot the extra afocal pickup... now I'm using wider aperture of 7mm; measured the exit pupil diameter and new estimate puts f/20 on the sensor. So before I was at f/40.

Traces of Coma and Spherical aberration start popping out, but not so much.

 

Turns out that the lines of a domestic neon tube are not narrow; here a "professional" image (=picked up from the net) of how Neon lamp spectrum should look like.

See comparable image at post #28.

.

.

... anyway, here a detail of the Neon spectrum from the radial spectrograph (a crop of full image like at post #27); looking at the width of the Red peak, resolution is a bit better than before, 10-20 pixels depending on taste. 

.

.

Spatial resolution a bit better as well, with separation of 20-pixel cycles

.

Edited by patta, 05 March 2025 - 09:00 AM.

[patta]

a fearless electrician friend helped me lit up some high voltage UV and Argon lamps to test the spectrometer.

 

Mercury vapor UV curing lamp

.

.

Argon glow starter:

.

.

Not so impressive, but something! It is approaching the resolution limits of the sensor.

there is a lot of longitudinal chromatic aberration (UV in focus, red not)

Edited by patta, 12 March 2025 - 02:18 PM.

[patta]

A day with clear sky! 

Rushed to mount the Lead Flint spectrometer on my advanced astrophotography rig

Vintage achromat 102-900  and alt-az AZ3 mount

 

Alignment is questionable

.

 

Counterweight and ND filter mount - go with tape

.

.

Otherwise it worked (didn't fall apart)

Edited by patta, 14 March 2025 - 07:21 AM.

[patta]

Here one of the images taken to the Sun with the arrangement of previous post.

 

It spells only troubles: the system is badly miscollimated; defects of the slit are visible (probably dents in the knife edge); the color sensor is playing tricks with the spectrum.

 

Good sides:

One quarter of the slit is usable

Good contrast, deep blacks.

The brass "cone slit" had no heat issues

Eye and camera not burnt

 

The system got dismantled after the sunny day and should undergo some fixes.

 

Sun limb spectrum, 13/03/2025

.

Edited by patta, 14 March 2025 - 08:31 AM.

[patta]

Further issue, the circular slit is very tight around the Sun disk, so with the poor AZ3 mount is nearly impossible to keep it centered around the Sun image.

.

.

One image off-center shows the border of the sun, with some decent contrast:

.

.

The best images came from lucky shots grazing near the edge, so there is a part of the slit really pointing at the chromosphere.

 

The light beyond the edge may be some low angle scatter or maybe actual Sun's limb.

Clear color bands like "flash spectrum" of the eclipse, are not visible; nor could I see prominences.

.

.

Needed improvements:

- centering 

- less weight and more stiffness

- find a friend with proper scope mount

- UV-pass filter

- new cone-slit

- monochrome camera

- need calibration lamp on the scope to focus and center the spectrograph

- ....

Edited by patta, 14 March 2025 - 07:26 AM.

[patta]

That's a fancy shot from earlier in the morning

Prominences? No, the sun was low on the horizon, we are looking at the branches of a tree....

.

 

Attached Thumbnails

•