Spectroheliograph: First light.

[Icosahedron]

Spent a frustrating afternoon yesterday chasing sunlight here in the South East. One of my USB to Cat5 converters stopped working the day before resulting in a rather taut USB cable out the back door. No way of observing and making manual adjustments as the instrument was designed purely for remote adjustment. Unable to remove dust from the slit and discovering that I didn't allow for long enough camera exposure time. Worst problem was the clouds scuppering my attempts to find spectral lines with an instrument that is not calibrated. By a miracle the sky cleared long enough to make an attempt. A long list of problems to attend to but in principle it works!

Brief description of spectroheliograph: Arduino DUE bit-banging a linear CCD sensor, positioning the optical components and handling data communication. Ordered a Teensy 4.1 last night. Did not have time to take photo's, that will come later.

 

[andrew s]

Good start.  Looking forward to a more detailed description of your instrument and further results. 

Regards Andrew 

 

[Merlin66]

Well done!

More info please.....

(your focus looks good!)

 

[skybadger]

Me too, let's see it!

 

[michael8554]

Intrigued, so looked up definition of Spectroheliograph.

Sounds like a Solar Ha scope, but you can alter the frequency of the pass band ?

Something like this ? 🙂

 

[Merlin66]

Michael,

That's what they were like in the early 1900s.

With the introduction of the webcam in the 1990s we were able to simplify the design and using digital scanning to assemble the solar image.

The current digital spectroheliograph (SHG) is very capable, and the grating in the spectrograph section can be tuned to any central wavelength allow us to image in CaK, Mg, He, Na, H beta, H alpha etc etc.

The software used to "reconstitute" the solar image has also improved over the years. Wah-Heung Yuen's "Spectra Line Merge" and John's "BASS Project" are among the freeware favourites.

My recent book "Imaging Sunlight - Using a digital Spectroheliograph" covers the subject, and describes many current SHG's and how to make and use them.

[Icosahedron]

Additional information:

Telescope: 80mm f10 achromat Barlowed to f22.5
Spectroscope: 2 x 900mm focal length achromats and 1800lpmm(30 x 30mm) grating
Camera: 2088 pixel(14 x 14µm) CCD line sensor with 16-bit imaging signal processor
Dimensions: 155 x 30 x 14cm
Weight: 11kg

[Merlin66]

Interesting.

I have a spreadsheet which may be of interest to others considering a SHG.

 

SimSpec SHG.xls

[michael8554]

Are "Camera" and "Image Sensor" two different items ?

Which parameter are you aiming for in "Summary" ?

Michael

[Icosahedron]

And finally...

[MarkAR]

Nice bit of kit, reminds me of the robot from Interstellar.

[Icosahedron]

28 minutes ago, MarkAR said:

Nice bit of kit, reminds me of the robot from Interstellar.

😁, thanks, now I can relax.

[Merlin66]

You may be interested to know that back in 1992 the very first digital spectroheliograph was conceived and built by Philippe Rousselle using a single line scanner (2048 x 1 array, 14 micron pixels) - almost identical to yours.

He used  an ATARI 1040 computer to interface and control the array.

 

Edited September 12, 2020 by Merlin66

[skybadger]

small then . I was expecting a shoebox on the side of a scope!  Did you build the entire mount to carry the SHG ? Or was it something you had left over from other builds ?

Will you be mountng this outside and imaging permanently ? 

I guess you use a panning mirror to scan the image across the slit or just use the sidereal motion ? sorry, bunch of questions. 

 

[Icosahedron]

6 hours ago, michael8554 said:

Are "Camera" and "Image Sensor" two different items ?

Michael

In the above diagram camera refers to a lens.

3 hours ago, skybadger said:

small then . I was expecting a shoebox on the side of a scope!  Did you build the entire mount to carry the SHG ? Or was it something you had left over from other builds ?

Will you be mountng this outside and imaging permanently ? 

I guess you use a panning mirror to scan the image across the slit or just use the sidereal motion ? sorry, bunch of questions. 

 

The mount is part of an abandoned project from years ago and it's very satisfying finally putting it to use. I'll be moving it in and out of the house as it's on wheels and fits through a standard door frame.

Operation will be as follows: The line scan camera runs continuously while the mount tracks the sun, allowing the wavelength and focus at that wavelength to be set. By analysing the scanned line (lit length and position within line), tracking can be corrected in RA and DEC. When instructed it then moves ahead of the sun, stops tracking and an image is acquired. Once calibrated it should be a fairly rapid automated process.

[Icosahedron]

Had another session with a clear sky and a gusty wind. The curve as seen here in the telluric lines indicates that the optical components are not in alignment. This misalignment becomes evident when viewing lines with a narrow bandwidth as can be seen further down. The image sensor requires leveling as the focus on the left differs from the right. Newton's rings can also be seen on the left.

The maximum camera exposure time limit has been increased in software and the slit assembly replaced with one that permits easy cleaning. This time I tried a slightly wider slit which has harmed resolution. I decided not to spend time resetting it and pressed on with other checks.

My next step was ascertaining whether focus could be reached at the other end of the spectrum. Then, thanks to the grating equation and interpolating focus settings, I made quick visits to H-beta and one of the Sodium lines. This was rapidly followed by checking H-alpha again to confirm that thin cloud wasn't rolling in. The curve in the H-beta and Sodium lines can clearly be seen. The lines also appear to have moved relative to the sensor in the Y-dimension. The structure is stiffer in the Y-dimension with a smaller area exposed to the wind and if the wind was causing it to flex I would expect it to be predominantly in the X-dimension as the wind gusts show. Totally perplexed at the moment. Any ideas?

[Merlin66]

You are achieving some excellent results. Well done.

The curved spectral lines "smile" is not uncommon, and is aggravated by the use of long slits (as used in the SHG) The rays passing from the extreme ends of the slit reach the grating at different angles from the central rays and consequently have different refractive angles. The curvature varies with wavelength.

In a SHG which uses a CCD/ CMOD detector this smile (and any residual tilt/ slant) can be corrected in the software before processing to an image.

With the linear array it may a case of compromise between the resolution and the impact of the smile.

Variation of the line position - this is usually associated with the grating positioning and the accuracy of the grating holder. Any differences between the axis of the grating grooves and the axis of rotation can cause some movement.  There may be the added complexity of thermal/ structural deflection etc. etc.

(Again, these issues are not as critical in the digital CCD/CMOS arrangement.)

Your eyesight must be better than mine

I don't see any significant Newton Rings, but if they are there, they are caused by interference within the cover plate of the sensor. In narrowband solar imaging they are worse with reduced bandwidth and increased focal ratios. The usual solution is to apply a slight tilt to the sensor.

Onwards and Upwards.

 

 

[Icosahedron]

Merlin66:

Thanks for the information. I'll start with checking the effect of the slit assembly temperature at the next opportunity.

[Merlin66]

[name],

I'm impressed with the performance you are achieving!

What height slit (and width) do you use?

It's all the more impressive as you are using a 30 x 30mm grating which I think would be vignetting the image.

Ken

 

Edited September 23, 2020 by Merlin66

[Icosahedron]

The slit length is 45mm, more than twice the required length as I find it easier to make than a short one. So far I've simply set the width to the minimum possible. The second set didn't quite match the quality of the first which is why I left it a bit wider.

Vignetting is present and fortunately can easily be removed during processing. Not contemplating upgrading the grating as yet.

Today I received the linear CCD array (5000 x 7µm x 7µm pixels) that I'd prefer to use. I don't expect image qaulity to improve as a result, but the spectral bandwidth will be halved. The imaging system will have to be sped up hence the mention of a Teensy in the excitement of my first post. The prototype camera board is a mess being subjected to experimentation. I've already prepared a new board that can accommodate either sensor via solder links.

[Rusted]

Very impressive set up, build and results!

You can tell that some time [50 years?] has passed since I used four safety, razor blades, four wooden clothes pegs.

This was for the slit. Two home made 60mm achromats in cardboard tubes on a triangular wooden frame made me a crude solar spectroscope.

I could see hundreds of Fraunhofer lines but not the H-a proms I had hoped for.  That took another 50+ years!

[Rusted]

On 12/09/2020 at 04:37, Merlin66 said:

Interesting.

I have a spreadsheet which may be of interest to others considering a SHG.

 

SimSpec SHG.xls 55 kB · 15 downloads

That is an amazingly narrow bandwidth! Is this typical of the resolution of the entire instrument?

[Merlin66]

Rusty,

Yes! When you want to record the Zeeman magnetic effects you need R>30000.

The capabilities of the digital SHG put to shame all the “commercial “ solar filters.

We can, as I said, image in any wavelength with resolutions down to <0.1A. How many expensive filters would you need to be able to do that?????

[Merlin66]

Icosahedron,

Just wondering if you managed to get the SHG to image the latest AR2779????

[Icosahedron]

No, I'm still not in a position to give it another go.

Over the last week I've built the new camera and got it working. Currently giving it a soak test while still waiting for a few non-critical components:

Next step is making a new slit. I'm now convinced that the movement of the lines was due to thermal convection currents inside the unit and will in future cool the slit assembly while observing:

I've abandoned the idea of replacing the Arduino DUE with a Teensy 4.1. Unless I'm mistaken Teensyduino does not support Arduino's SerialUSB object and their Serial object is fixed at full USB speed. In this application the DUE outperforms the Teensy.