Spectroheliograph.. my version

[bojan]

Inspired by Ken's implementation (https://www.iceinspace.com.au/forum/showthread.php?t=193393) of Christian Buil's design (http://www.astrosurf.com/solex/sol-ex-presentation-en.html) some time ago, this is what I have done so far... No images yet, just ideas on how to put together SHG with what I have on my junk yard...

The biggest challenge is slit (image of 1/2 blade through optical train is attached), considering it has to be very narrow (10 um as per Christian Buil), but I think I solved it - at least in principle, using Gillette razor blades and simple sliding mechanism to adjust the slit width. Hopefully it will work. I even have miniature stepper with linear actuator, I may use it at some stage in the future if I am happy with operation of this instrument.

Optical elements are: old Industar 50 lens as colimator, foil grating (1000 l/mm) and old Kyocera M400R camera with it's 10x zoom lens.

Everything is mounted on improvised frame made from cut-offs of "L" 50x50x3 mm extrusions (with machined holes for optics), assembled and tightened with screws...

[Merlin66]

I admire your work…

Pencil sharpener blades, with the single bevel work better than razor blades.
It’s more usual to work with higher resolution reflection gratings.
For the SHG the minimum seems to be around 1800 l/mm with the standard at 2400 l/mm
The Camera really needs to be fast frame, capable of >60 fps and a typical solar scan will take up to 3 Gb depending on the chip.
The INTI program to convert the AVI to a quality image has undergone some significant improvements in the last few months and the latest V3.5 is very capable.
Onwards and Upwards.

[bojan]

Hi Ken,

Thank you for the comments..
This is just my attempt to re-purpose what I have, and by no means the last one.. If I find observation with this instrument interesting (for me), I will invest in more capable camera and better optics.. This camera has 30 f/sec, and supports AVI format.

[bojan]

This arrangement of the slit width semi-adjustable block is better, I believe...
The left side (with 1/2 of the blade) is screwed to the bottom of the adapter, the right block (with second half of the blade) can be moved up-down to adjust the slit width.
It is easy to add screw for manual/motorized adjustment as well.
I will add the springy shim to hold the right-hand block (with other blade) down to the base plate (using two holes for screws on the left).


And all this is easily machinable...

[Merlin66]

This arrangement has been used on other spectroscopes.
Just check the wedge angle vs closure.
Pencil sharpener blades are still better.

[bojan]

Yep.. I can still use pencil sharpeners down the track, the same principle applies...
Angle I will use (hopefully) will be 1°, this will give me fine enough adjustment.

BTW, I ordered them from ebay (https://www.ebay.com.au/itm/193637786007) just now.

[bojan]

First spectrum.. Energy saving lamp.
I will definitely need better grating (more l/mm) and more narrow slit (this one is made from 2 razor blade halfs.

[bojan]

After watching this video (https://www.youtube.com/watch?v=h29NsNGDOw8), I decided to try this approach, using pencil sharpener blades.
My implementation should look something like in the attached sketch.

Also, it occurred to me I could try to use here my Hartridge spectrometer..
There will be some limitations (very narrow FOV), now waiting for neutral filter for my Rubinar 8/500 (to avoid possible thermal damage to parts close to focal point).

[Merlin66]

Rupert has been playing with various slit designs for many years…….not always successfully.

[hamishbarker]

Have you considered using a reflective cylinder as a slit?

The effective slit width is dcyl / (4* focal ratio of input telescope) * cos (theta/2)

where d cyl is the reflective cylinder diameter, theta is the angle between input and output.

For example, f10 input and a say 1mm diameter needle with 90 degree direction change, effective slit = 1mm/40 *cos 45= 18 micron.

Google Maurice Gavin's various papers on the subject. He successfully built various SHGs based on the technique.

[bojan]

Good idea, thanks..
Actually, I was using small bearing ball as artificial point source once long time ago.. same principle.


So, the concept should look like on below sketch..

[bojan]

Quick improvisation... It looks reflective slit is better.

[bojan]

Solar spectrum from hand... Magnesium triplet is clearly visible.
Needle (reflective slit) has some issues, (not polished properly.. I need to find a better one) but generally resolution is decent.
Also there is a lot of stray light.. I need to think of some sort of enclosure for the contraption.

[Merlin66]

Bojan/ Hamish,
I spent some time with Maurice when I stayed in Cobham. He used to visit me and I'd go over to his observatory (12" Meade).
We did discuss, among other things, various slit designs, including reflective needles.
The generally poor surface accuracy gives lots of Transversalium and the reflected beam spreads out over a wide arc - very inefficient. High speed rotation of the needle seemed to give slightly better results.
(See p142 "Astronomical Spectroscopy for Amateurs.)
I used a needle reflection to show the solar results which can be obtained with a grating and telelens.

[bojan]

Ken,

I wonder if glass fiber (ø0.12 mm) could be used? The surface appears to be much better, compared to the needles I have at the moment.
Since there is a lot (too much actually) of light from Sun, losses are actually welcome.

I will try it next week.

[bojan]

BTW, would it be possible to aluminize that fiber?

Asking this in case transparent glass does not produce good contrast?

[Merlin66]

Bojan,
There a guy in the US spray painting his large mirrors rather than aluminium vacuum coating. The details must be somewhere.....cheaper solution to test?

[SteveInNZ]

There's Angel Gilding spray silvering that's not really spray painting. It uses spray bottles but is really a chemical deposition.

Have a look in the ATM section and see if anyone is silver coating their own mirrors.

The optical fibre sounds like a good idea, even without a coating. It will certainly be smooth and it's designed to be as reflective as possible within the fibre so it would seem reasonable that the outside would be too.


Steve.

[bojan]

Guys, thank you for the tips.
If all goes as I hope it would, I could try this spray (prohibitively expensive!!), or better still, I could ask someone to put my fiber with their mirror in their aluminizing chamber...

[bojan]

Holman 25 mm PVC plain Tee (https://www.bunnings.com.au/holman-25mm-press-pvc-plain-tee_p3142350) is perfect for the purpose.. machined (with holes for precise positioning of the reflective slit) and ready for spray of black paint.
The tee-shape will also help keeping the stray light away from camera (to further improve attenuation, I may add some more black painted tubing at the end)

[bojan]

Reflective slit with optical fiber is excellent (no transversaliums).. but a bit dark, so aluminization should be in order. Also I wander if ghost lines are visible (left, internal reflections in fiber?)

On image below I captured green line (577.7 nm) from spiral lamp.
The whole thing is too fiddly for solar spectrum.. so I need to do some more of mechanical work.

[bojan]

Based on slit image (first pic from previous post) and measured fiber diameter (0.12 mm, 40 pixels), effective slit width is ~6 pixels, ~18 um.
Please note, this measurement is done without any optics.. and the scope I intend to use will be Rubinar 8/500 or Samyang 500/ f6.3.

And, if higher f/number for colimator (Industar 50) is used (currently 3.5, wide open), the effective slit width could be reduced to <5 um.
Not bad for the price I paid for it :-)

[bojan]

Thanks to Stefan, couple of optical fibers are aluminized to hopefully improve the reflective slit performance..

I sort-of assembled the whole thing on my workbench...
Now I am waiting for neutral filter (1/8) from alibaba.

[bojan]

Images of reflective slit before and after aluminization...
There is improvement in contrast and reduction of internal reflection..

I probably have to re-adjust the angle of the fiber to improve the illumination of the aluminized side.

[hamishbarker]

I've not been following the thread. Let me know if you need more/various fibers/glass rods (or mirrors but then post to/from nz may be a pain) aluminized. I have a working aluminizing chamber.

aluminizing a fiber may require at least 2 passes if a full coverage is desired in order to avoid the need to care about which side of the fiber you have facing your light source. The aluminium atoms fly in a straight line inside the chamber.

[bojan]

Hi Hamish,
My apologies for late response and thank you for your offer...
At the moment I have all the help I need from Stefan :-)

After quite some time I had a chance to have a brief look (visual, no camera yet) through my spectrometer, mounted on Rubinar 8/500.
I tried with camera, but there is too much stray light, so I need some sort of a black cloth enclosure to be able to take usable images.

Contrast and resolution are quite decent (there was no NF, still waiting for it)).
I managed to rotate the fiber in right direction, but still there are a lots of transversals.. so I wonder if we will be able at all to solve this problem with fiber.. Of course the standard slit approach is still the viable and tested option.
Anyway, this is a progress.. small and slow but still a progress..

[bojan]

Since my ø76 mm NF8 arrived couple of days ago, today I attempted another approach - with Hartridge spectrometer.
I already had necessary adapters ready, so it was just a matter of putting the system together in 10 minutes and try..
Spectrum looked OK, but Sun's image focus on slit was way off..
Then the clouds rolled in.
I suppose the challenge will be to control the camera exposure (I was using Free2X Webcam capture app, it seems it was set to Auto - I have to fugure how to adjust that yet).

Sun's image diameter may need to be scaled properly (at the moment, it was ~5mm in diameter (scope is Rubinar 500mm f/8, without Barlow).
Also, I should increase the spectrum scale, secondary band is not necessary and may confuse Sol'ex software.

[Merlin66]

Bojan,
The INTI software for the reconstruction of the solar image from your SHG relies on selecting an ROI which just shows the target spectral line.
With my camera, ASI 178 this means processing an SER video file of 3000 x 100 pixel. The SER file is usually about 2000 frames to give good coverage of the solar disk.
Ken

[bojan]

Ken, thank for the tip, I was hoping that is the case.

I am still at basic level with all this, but I feel this is the way to go for now.

[bojan]

I found camera adjustments, everything is at hand.
And clouds are still here.

[Merlin66]

Bojan,
I don’t see any frame rate control.
To get a usable SER file you need to be able to run at a high frame rate.
When you’re ready just drop me a PM or email.

[bojan]

Frame rate can be set to 30 f/sec max.
Once I sort out the electronics f/w for my tracker (I will set it to 10x faster than tracking), I will have ~18 sec for Sun disk scan (540 frames.. should be enough for the proof of concept).
Also, I have two cameras - Logitech 4000 Pro (640x480) and IMX291 (1920x1080).

The latter one has IR filter glued to the sensor cover glass, so I have to figure out how to remove it without damaging the sensor. Maybe with hot air soldering iron and scalpel to lift the filter.

[Merlin66]

Bojan,
Ok
I see you will take AVI files. Valerie’s INTI program only works with SER files.
The Smiths version works with both SER and AVI files.

[bojan]

There are AVI -> SER converters available.. this one for example:

https://www.converthelper.net/dv-avi-to-ser
SIRIL (https://free-astro.org/index.php/SER) can also do it.

[bojan]

I managed to obtain sensible video scan this morning in IR part (starts as narrow transversal line, gets wider and wider, then after a while starts to shrink until it disappears as Sun disk is traveling over the spectrometer slit), but apparently the focus is still not quite OK.
The mount I was using for this experiment (EQ3-2) has limited controls (only tracking), and unfortunately my LXD75 in at Mount Pleasant.. so the adjustment of the whole setup was fiddly, and sunlight glare on the lap screen didn't help either.
Anyway, now I have first-hand experience about what else needs to be done to make this procedure more comfortable (proper mount under full remote control, sun shade for lap, comfortable chair etc..).


The fat line on the spectrum below is at ~750 nm (according to dial on the spectrometer - I am not sure how accurate it is in IR since it is calibrated for visual), most likely O2 band (760 nm)

And yes, I have to clean the sensor...

[Merlin66]

Bojan,
Yes, That looks like the Telluric line at 7400A.You should be able to use the scale to get pretty close to H alpha at 6563A


Onwards and Upwards
Ken

[bojan]

Yes, I was going for the most prominent feature in the spectrum.
Dial is actually quite accurate, even in IR.
This photo was taken with modified Logitech 4000 Pro, and "blue" is not blue, is is IR.
I may try again H-alpha later today or tomorrow.

[Merlin66]

Bojan,
Using an ROI to isolate the immediate area around the target line will minimise the size of the capture file. They can get pretty large!!
Many SHG users use FireCapture (freeware) or SharpCap (pay ware) to control the camera and ROI
I use FireCapture.
When the solar disk is in focus, the top and bottom edges of the spectrum should be clean and tight.
The next step then then to get best focus WITHIN the target line - that’s where the data lies to for the SHG image.
You’ll start to see wiggles bright and dark spots within the line as you get close to usable focus.
Ken

[bojan]

Ken,

Thank for the tips..This was roughly what I was doing...
Focus of the Solar disk is not easy to achieve, the focus ring on that Rubinar (8/500) is quite tight.

I think I will try to come up with some sort of focusing screen, to get it close enough for the start. F/8 will be forgiving for couple of 1/10 of mm difference in slit distance from lens focal plane (I hope..)..
The capture software (Free2X Webcam) is quite easy to use, and camera driver allows for adjusting exposure time (minimum is 1/10,000, just fast enough. Slit is also close to minimum width, I can't go much lower in order to avoid transversals).
I tried couple of codecs, MP4 results in lowest compression artifacts. Maybe video camera without compression is the way to go (that will be in the future).

Attached is one frame from AVI file, southern (or is is northern?) solar limb is nicely visible. Obviously, 5 mm solar image diameter is too large for Hartridge slit, so I will have to do couple of scans to capture the whole disk.
Anyway.. slowly getting there.

[bojan]

I managed to convert AVI into SER format using PIPP (3 GB!).
Now, time for INTI.
(and I cleaned the sensor...)

[bojan]

Hmm..

INTI simply disappears when I press "OK"..

[Merlin66]

Bojan,
I assume you’re looking at Valerie’s latest version V4 (I think)
It does take some time for the first load….it has to install thePython code etc.

[bojan]

I know about that delay..
This is something else. I went into directory where SER file is, and I selected it, then pressed OK.

And that was it - the whole thing disappears.

BTW, I had latest python installed already.

[bojan]

OK, I think I know.. my Py version (2.7.14) is not *the* latest, and latest doesn't run on W'7.

[Merlin66]

Bojan,
You could also look at the Smiths version…
https://groups.io/g/Solex-English/message/311

I’m not at home at the moment so can’t help much.
We should be back home tomorrow.
Remember the processing software INTI etc expects to see one one absorption line in the video. It then selects this line to process into the SHG image.

[Merlin66]

If you can’t control a capture ROI with the camera.
ImageJ can crop AVI files. This may be a way around…..

[bojan]

Yes..
Free2x Screen Recorder can do the same.
I tried with Python 3.7, same problem.
No worries, I will try w'10.

[Merlin66]

Ok
Did you select an ROI across the H alpha line for your trials?

[bojan]

No, not yet.
Free2X screen recorder has limitations (can't go narrower than 300 pxls.

PIPP can do video cropping (apart from conversion into SER), but ROI can be selected only numerically as pixel offset and width (not possible to select on screen via GUI).. not very user friendly for this job.


Also, I think I will need to use my other camera (IMX291-based, but first I have to remove glued IR filter.. for this surgery I need to be relaxed and right now I have to do other things (work from home).

So.. no rush, will try again by end this week or next.

[bojan]

OK, I know what happened.

On my slower w'10 lap, I was able to see message on CMD screen just before closing down:
"Acquisition not done at 16 bits"
On my desktop, I saw only flicker..

So - I need 16 bits SER file, not RGB 8 bits.

[Merlin66]

The Smiths version handles 8bit AVI and SER as well as 16 bit SER.

[bojan]

I tried IMX291 camera.. it is possible to get 10-bit depth, and resolution is much better, 1920x1080. But no exposure control..
So, I will try to enlarge the image with Logitech.. it is basically eyepiece projection using lens from old scanner so moving that lens closer to spectrometer and moving camera away will do the trick.
But it seems FireCapture doesn't recognize any type of (cheap) webcams.. so this is a show-stopper for now.

[bojan]

Ken, Where can I get this app?

OK, I think I found it (https://groups.io/g/Solex-English/topic/93850388?p=Created,,,20,1,0,0::rece ntpostdate/sticky,,,20,2,0,93850388,previd=166 6149402230337254,nextid=16641252895 72131018)...
But Avast doesn't like it.


However, source (python version) seems to be working..
Nope. Failed to install.

I am getting tired of this...

[Merlin66]

Bojan,
I've downloaded that file (windows version - you don't need the Python version)
Works 100% - no virus issues.

[bojan]

Hmm..

I will try to disable Avast... but not sure if that's a good idea.

[bojan]

OK, I downloaded w' version remotely from my desktop in the office (company uses Sophos Endpoint Engine for protection, and all went well, so Avast generated false alarm (not the first time..).
However, Avast keeps complaining about Trojan infection..

So, I don't know what to do here..

I think I will wait until Avast resolves this issue properly.

[Merlin66]

I hope you can finally resolve your false positive issue…..

[bojan]

It seems the only way I can resolve this for me & here is to totally disable Avast (it has a multi-level protection it seems, I tried to add exceptions at various pionts but to no avail), and I really don't want to do this..
I noticed Valerie mentioned in couple of places there are issues with INTI and antivirus apps..

[bojan]

While waiting for Avast false(?) alarm issue resolution, I tried SharpCap (v4.0), in hope for user friendly ROI selection.
It seems my camera (Logitech 4000 Pro, 640x480) doesn't support ROI?

Same with IMX291 camera..

[Merlin66]

A lot of SHG users use small pixel mono cameras similar to ASI 178.

[bojan]

With the help from theSmiths, I have SHG (python version) running on my w'7 desktop.

The next step is ROI extraction from video file I obtained yesterday and hopefully my first Sun image in H-alpha :-)

[bojan]

Well, SHG works, but it seems my data are not of sufficient quality..
First, PIPP ROI selection has problems, there are errors when I specify tall and narrow crop area near the middle of the original frame.
I tried with default frame size which contains a lot of rubbish (noise, inverse band (produced by spectrometer - I will have to screen that somehow if I can't find something better for ROI extraction)), so SHG is totally confused.
It seems SHG expects only one line in video, or locks on the first one found?
If so, perhaps I can position the line of interest (H-alpha.. at the moment it is near the centre of the frame) at the left-most side of the frame, in which case PIPP will be able to discard unwanted right part.

[Merlin66]

ImageJ should allow you to crop any area/ zone of your AVI file.
Yes, as I mentioned before INTI needs to see a prominent absorption line within the ROI.
I think you’re beginning to appreciate the issues involved in obtaining a worthwhile SHG image.
Your focus is probably not good enough to give a clean contrasty solar image.
Practice, practice, practice…….

[bojan]

Yep, learning from our own mistakes is the best way of learning, my mom used to say :)

[bojan]

Back to mechanicals...

Since using my EQ3-2 mount for the Sun scan takes 3 minutes (when tracking is disabled), I searched for a quick solution for suitable tracker.. and found this one (https://www.thingiverse.com/thing:2422008) (The first one, tracker).

The only variables I needed to modify were TIM_FAST. and char sm_tab[8][4] (I am using full step with pause to save power).
The rest was OK for my mechanical setup.

New electronics (Arduino NANO) will fit inside EQ3-2, inserted into existing PCB (designed for 68HC705KICS (pic attached), which I still intend to use for that mount).

[bojan]

I found the old objective from Russian Smena-8 35mmm film camera in my stash..
FL is 40mm, outside diameter is a little bit too wide (ø24 mm) but it will fit inside barrel-adapter as projecting lens onto IMX291 sensor (I managed to remove the IR filter, it was glued with silicone rubber on the glass lid).

[bojan]

IMX291-based camera (https://www.iceinspace.com.au/forum/showthread.php?t=194620&highlight=IMX291&page=2) mounted on Hartridge spectrometer (tuned to 656.3 nm), ready for test... (clouds.. everywhere).

Adapters were machined from PVC 32mm fittings from Bunnings.
Shown on the second image is the spectrum of spiral fluorescent lamp, taken with the same setup.

[bojan]

I gave up on Hartridge spectrometer (for now). Mostly because of focusing (not easy at all) and relatively low aperture (effective slit length is ~2mm, maybe a bit more).
So I am back to fiber reflective slit...

In the meantime I figured out the Holman T32mm tee joint is mechanically perfect for improvised housing for fiber slit.. It can accommodate M42x1 (or T2) threads with minimal machining (ø42mm, so I have to add ~0.6 mm of material in a form of a thin ferrule, or use one of those (https://www.ebay.com.au/sch/i.html?_from=R40&_trksid=p2334524.m570.l1313&_nkw=M42+adapter&_sacat=0&LH_TitleDesc=0&_odkw=M42+&_osacat=0&_sop=15) from ebay), so Pentax (or Macro Rubinar 500mm F/8 in my case) lenses can be easily attached to the setup

Also, I added eyepiece to project Sun's image on white screen (together with fiber) to make focusing on slit easier.

So - there will be another version of the hardware, with some 3D printed parts (articulate joint to select wavelength)...
Attached Thumbnails

[bojan]

A bit better design for 3D printed parts..

[bojan]

Camera tube is finished (~1 hour) after some troubles with 3D printer (I think I stumbled on a bug in app.. not sure yet).

[Bassnut]

mmm, layers look a tad rough, 0.1mm?

[bojan]

Actually, 0.2mm. There is no need for anything finer, I will machine the part to more precise fit anyway...

[bojan]

Printing started beautifully.. but after 6 hours - disaster.

[kencas]

PET-G? I have just finished printing the Sol'Ex SHG and I found when I was printing the main body parts that I had to turn off "Part Cooling" in the slicer program because I was getting layer non-adhesion, cracking and warping. It might be worth a try to turn it off with a re-slice of the STL?

[bojan]

it is ABS..

Yes, I was thinking about cooling, but then again switching it OFF may affect supports..

Upon closer inspection I saw the problem probably started with poor adhesion of the base to the plate at edges.. ((maybe plate was not clean enough?)

So next time I will try with increased filament flow a bit (by 1%, to improve layer sticking) and to change the design by separating base and walls into smaller segments and glue them together later (and save some filament in case individual parts are usable or repairable).

As for this particular print session, I removed separated layers and figured out I can re-use the base, since the separation occurred abruptly at 11 mm height. So I will just print what's missing and glue it together and see how it goes.

[bojan]

Again.. 1.5 hours to go. Not as bad, this part will be usable. I will fill the gap with slush.

[TrevorW]

Great work, ABS always have issues with adhesion try using a raft, I mainly print in PLA made a couple of camera adapters without any issues strength wise :)

[bojan]

Here is a result... now, gluing (with acetone). And perfect fit... as per design :-)
For adhesion I am using capton tape.. usually it is OK, but it needs to be cleaned with alcohol immediately before print job.. Sometimes it is hard to remove the finished part.

Then I will be printing inner ring..

[bojan]

Inner shell (that will hold Holman tee and Smena-8 lens) is finished..
There was a bit of layer splitting, but I applied superglue to the crack when noticed it is happening..
The next step now is printing the base lid, then gluing to the shell, and then machining to have a fine fit when rotating shells inside each other.

[bojan]

Machining and adding threads to tee 32mm (cheap M42 & 39 mm adapters from ebay)...

[bojan]

I tried to assemble the housing after gluing the individual parts, the whole thing fits better than I hoped for.
Insides of 32mm Tee adapter is painted black, top lid (with adjustable grating bracket) is fitted, aluminized fiber inside Tee is still missing (not easy to mount it, it is quite fragile and easy to break... this operation I left for the very end, it can be done only with the help of the stereo microscope and very fine tweezers).

[bojan]

After a long delay (too long !!!) I am back to this..
I managed to insert the very thin fiber (not sure of the diameter, it came from optical cable bundle, donated by Stefan), and gave it a go...
There are still a lots of transversals (dust on fiber!) but judging from the Sun's limb (last image, lower part), overall quality of recording is much better with new version (still a long way away from the current standards..) but I am hopeful this concept will work good enough one day.
Second image is Sun finder.. simple eyepiece projection of the fiber and Sun image at the focal plane of the Rubinar 8/500 Makro. It makes the focusing much easier.

[bojan]

It just occurred to me today, that I could try to solve the transversals issue by cleaning the fiber (even after mounting in place) in the ultrasonic bath.. I will try on first occasion and report.