Star Analyzer Prism Myths


Tim Stone
 

There seems to be a significant uptick lately of interest in the Star Analyzer Prism. This prism is an accessory designed specifically for use with the SA100 grating. Because I had some conceptions of the purpose of this prism that turned out to be simply and utterly incorrect, I thought I might relate those conceptions and hopefully clarify what the prism is actually for.

When I decided to make the jump, I had to decide what hardware I wanted to use. I didn't have a grating, and so I needed to buy one. I bought the SA200, under the basic notion "If 100 is good, 200 is better." This not entirely true. It's similar to saying if a little screwdriver is good, a big old honking screwdriver is better. Like the screwdrivers, the SA100 and 200 are both good tools, and useful for what they are good for. It turns out, the 100 would have been a better learning grating for me. If you're just venturing into this space, I would recommend starting with the 100, and then adding the 200 if or when you decide you want to do something that tool is better for. But that's not what this post is about.

I also bought the 3.8 degree prism. I thought it magically increased the resolution I would attain with my SA200. I was giddy with the thought that I could simply screw a wedge-shaped piece of glass onto my grating, and get even -better- results. At first, I acquired some spectra with my SA200 without the prism. I can tell you I was so thrilled to see absorption features, I could hardly contain myself. Imagine my thrill at the thought of doing EVEN BETTER than that with my prism. I screwed it on, acquired a spectrum from one of my previous targets, and... I didn't see any difference at all.

I thought I'd been ripped off. The thing clearly didn't do what I thought it would. Why did I spend all those $$ for it? In my disillusionment, I put it away and just concentrated on using the grating. That was the first correct thing I did.

Now, a couple of years later, I have come to understand the purpose of the prism, what it does, and what it does not do. Let's start there: what does it not do?

  1. It does not significantly improve your dispersion or resolution.
  2. It does not necessarily make your spectra better. I say -necessarily- because it -can- help in one way, I'll talk about in my list of what it -does- do.
  3. It is no substitute for quality data acquisition (e.g. focus, exposure), processing (e.g. stacking), and reduction (e.g. calibration) skills.
Let me dwell on point #3 for a minute. To even -see- the effect of the prism, you must be able to attain good focus. Focus is the #1 problem I had with my spectra. Even when I thought for sure I was in focus, I was way out of focus. I was chronically out of focus, no matter how little or much spacing I had between my prism and my sensor. If you don't know how to get excellent focus, then that's the only thing you should be working to perfect. Everything depends on that. (Hint: when the spectrum strip is in focus, the zero order star is not.)

Now, what -does- the prism do?

  1. It makes the spectrum strip distinctly non-linear. This is how prisms work. This introduces a lot of bother in the process of calibrating your spectra. You can no longer calibrate on the zero order and a well defined line. It won't work that way anymore. You will need to do at least a three-point calibration. If you're not really good at calibration, stay away from the prism. Once you can get excellent focus, then concentrate on becoming great at calibrating on two points. Graduate from needing the zero order as one of your points, to using two lines, one at the bluer end and and one at the redder end, to do your calibrations. Get good at this.
  2. For the bother (should I say "pain?") of non-linear calibration, you get a reduction in aberrations at a specific location in your spectrum image. The prism moves the spectrum strip toward the center axis of the optical system, where theoretically, aberrations such as coma and chromatic are null. The wavelength where the aberrations are null is determined by the grating and the prism deviation angle. For the SA 3.8 degree prism, the deviation angle is about two degrees. With the SA100 and the prism, at the grating/sensor spacing they were designed for, the aberration null wavelength should be about 6500A. With my SA200 and the prism at the spacing I use, the null is around 4300A. 



Once you've learned to acquire excellent data, process it well, and reduce (calibrate, normalize, etc.) it accurately, perhaps it is time to add a prism to your mix. It seems the null wavelength shifts with your grating/sensor spacing, so you will have to experiment some to position the wavelength you want to study at the null point. While you will not really have any higher resolution than without the prism, you will have lower aberration around the null point, and so your results there will likely be closer to the maximum for your system. The difference, though, is not nearly as dramatic as you might think.

Having been deeply involved in deep-sky astrophotography for years, I had gotten to the point where little things mattered to me. I was fine-tuning my deep-sky images in ways only I could see. When I stepped into spectrography, I had forgotten what it was like to be way down low on the learning curve, and naturally thought I understood everything well enough to continue to care about the little nitpicky things. It took me most of the first year to finally come to the realization that I didn't know anything about what I was doing, and needed to just go back to the fundamentals everyone here on this forum was trying to tell me about:

  • Practice, practice, practice
  • Learn to get awesome focus
  • Learn to get exposure correct
  • Learn to calibrate
  • Stick with bright stars
  • Learn with A stars first. 
  • Learn the cooler stars after the A stars.
  • Start tuning your spacing to see if you can get better results
  • Do some awesome stars like P-Cyg, a little dimmer, but still reachable
  • Do some Be stars or Mira variables
  • Now, are aberrations bothering you? Add a prism.

Last, let me offer my apology if this seems patronizing or pompous.I don't intend any offense. I don't know it all. I'm still only two years into this avocation. I may have some details wrong in this post. If so, I hope the forum will correct me! But... I'd just hate it if people had the same misconceptions I had about the prism and, like me, made decisions based on those misconceptions. If you're new, just know, there is a place for the prism, but you don't need it to take fantastic spectra!.


Ken Harrison
 

The grism (gratings-prism) was proposed by Edwin Carpenter back in 1963.
The Carpenter prism has the front face square to the incoming light beam  (no deviation or dispersion), at the rear prism face a transmission grating is mounted (cemented) and as the light exits the prism there is deviation but no dispersion. The angle of the prism and it's deviation angle is designed to bring the spectral wavelength of interest back onto the optical axis.
This makes the mechanical and optical design of the spectrograph much easier. All professional transmission grating spectrographs make use of the Carpenter Prism. (I assume the ALPY also uses the Carpenter prism.)
Edwin F Carpenter at Steward Observatory describes the "grism" constructed and used on the 36" telescope in 1963.http://articles.adsabs.harvard.edu/full/1963AJ.....68R.275C
See Daniel Schroeder "Astronomical Optics", Chapter 15, Plane grating Spectrometers, Section iii NonObjective grism, He describes how the inherent coma and defocus at the ends of the range (field curvature) can be significantly improved.
The following patent describes the Carpenter prism at some length but not the source of the name....
https://www.google.com.tr/patents/US20090034077
--
"Astronomical Spectroscopy - The Final Frontier" - to boldly go where few amateurs have gone before....


conrad cardano
 

Hi

There is some interesting viewpoints here.  Let me add some of my observations to it.

I use Rspec to help me with the focus.  It allows me to measure just how deep the aborption lines are.  The deeper, the better.

I have been using the Star Analyzer 100 for over 8 years.   I have it attached to a refractor with a dual speed focuser.  These focusers are common and allow for 1:10 ratio in fine focus.  A quarter turn in the micro focus can make the difference between a good spectra and a great one.  Whenever I have conected it to a DLSR, I find it very difficult to get a precise focus.

Lastly,  the quality of your optics makes a huge difference.  I have a 3” f/6 refractor with ED optics.  It does a good job focusing light in the visible range.  Don’t forget, that’s what it was designed to do.  It does not do a good job with the focus in the deep blue or near IR.  No one can see that light.  It is common for deep sky photographers to insert UV and IR filters in the light path to exclude that light.  A reflector does not have that problem; although short focus reflectors have their own optical abberations.

Conrad Cardano
Rhode Island, USA


On Thu, Aug 27, 2020 at 10:13 PM Tim Stone <tim.stone.piano@...> wrote:
There seems to be a significant uptick lately of interest in the Star Analyzer Prism. This prism is an accessory, designed specifically for use with the SA100 grating. Because I had some conceptions of the purpose of this prism that turned out to be simply and utterly incorrect, I thought I might relate those conceptions and hopefully clarify what the prism is actually for.

When I decided to make the jump, I had to decide what hardware I wanted to use. I didn't have a grating, and so I needed to buy one. I bought the SA200, under the basic notion "If 100 is good, 200 is better." This not entirely true. It's similar to saying if a little screwdriver is good, a big old honking screwdriver is better. Like the screwdrivers, the SA100 and 200 are both good tools, and useful for what they are good for. It turns out, the 100 would have been a better learning grating for me. If you're just venturing into this space, I would recommend starting with the 100, and then adding the 200 if or when you decide you want to do something that tool is better for. But that's not what this post is about.

I also bought the 3.8 degree prism. I thought it magically increased the resolution I would attain with my SA200. I was giddy with the thought that I could simply screw a wedge-shaped piece of glass onto my grating, and get even -better- results. At first, I acquired some spectra with my SA200 without the prism. I can tell you I was so thrilled to see absorption features, I could hardly contain myself. Imagine my thrill at the thought of doing EVEN BETTER than that with my prism. I screwed it on, acquired a spectrum from one of my previous targets, and... I didn't see any difference at all.

I thought I'd been ripped off. The thing clearly didn't do what I thought it would. Why did I spend all those $$ for it? In my disillusionment, I put it away and just concentrated on using the grating. That was the first correct thing I did.

Now, a couple of years later, I have come to understand the purpose of the prism, what it does, and what it does not do. Let's start there: what does it not do?





  1. It does not significantly improve your dispersion or resolution.


  2. It does not necessarily make your spectra better. I say -necessarily- because it -can- help in one way, I'll talk about in my list of what it -does- do.


  3. It is no substitute for quality data acquisition (e.g. focus, exposure), processing (e.g. stacking), and reduction (e.g. calibration) skills.




Let me dwell on point #3 for a minute. To even -see- the effect of the prism, you must be able to attain good focus. Focus is the #1 problem I had with my spectra. Even when I thought for sure I was in focus, I was way out of focus. I was chronically out of focus, no matter how little or much spacing I had between my prism and my sensor. If you don't know how to get excellent focus, then that's the only thing you should be working to perfect. Everything depends on that. (Hint: when the spectrum strip is in focus, the zero order star is not.)

Now, what -does- the prism do?





  1. It makes the spectrum strip distinctly non-linear. This is how prisms work. This introduces a lot of bother in the process of calibrating your spectra. You can no longer calibrate on the zero order and a well defined line. It won't work that way anymore. You will need to do at least a three-point calibration. If you're not really good at calibration, stay away from the prism. Once you can get excellent focus, then concentrate on becoming great at calibrating on two points. Graduate from needing the zero order as one of your points, to using two lines, one at the bluer end and and one at the redder end, to do your calibrations. Get good at this.


  2. For the bother (should I say "pain?") of non-linear calibration, you get a reduction in aberrations at a specific location in your spectrum image. The prism moves the spectrum strip toward the center axis of the optical system, where theoretically, aberrations such as coma and chromatic are null. The wavelength where the aberrations are null is determined by the grating and the prism deviation angle. For the SA 3.8 degree prism, the deviation angle is about two degrees. With the SA100 and the prism, at the grating/sensor spacing they were designed for, the aberration null wavelength should be about 6500A. With my SA200 and the prism at the spacing I use, the null is around 4300A. 







Once you've learned to acquire excellent data, process it well, and reduce (calibrate, normalize, etc.) it accurately, perhaps it is time to add a prism to your mix. It seems the null wavelength shifts with your grating/sensor spacing, so you will have to experiment some to position the wavelength you want to study at the null point. While you will not really have any higher resolution than without the prism, you will have lower aberration around the null point, and so your results there will likely be closer to the maximum for your system. The difference, though, is not nearly as dramatic as you might think.

Having been deeply involved in deep-sky astrophotography for years, I had gotten to the point where little things mattered to me. I was fine-tuning my deep-sky images in ways only I could see. When I stepped into spectrography, I had forgotten what it was like to be way down low on the learning curve, and naturally thought I understood everything well enough to continue to care about the little nitpicky things. It took me most of the first year to finally come to the realization that I didn't know anything about what I was doing, and needed to just go back to the fundamentals everyone here on this forum was trying to tell me about:





  • Practice, practice, practice


  • Learn to get awesome focus


  • Learn to get exposure correct


  • Learn to calibrate


  • Stick with bright stars


  • Learn with A stars first. 


  • Learn the cooler stars after the A stars.


  • Start tuning your spacing to see if you can get better results


  • Do some awesome stars like P-Cyg, a little dimmer, but still reachable


  • Do some Be stars or Mira variables


  • Now, is coma bothering you? Add a prism.





Last, let me offer my apology if this seems patronizing or pompous.I don't intend any offense. I don't know it all. I'm still only two years into this avocation. I may have some details wrong in this post. If so, I hope the forum will correct me! But... I'd just hate it if people had the same misconceptions I had about the prism and, like me, made decisions based on those misconceptions. If you're new, just know, there is a place for the prism, but you don't need it to take fantastic spectra!.









R L
 

On Fri, Aug 28, 2020 at 04:59 AM, Ken Harrison wrote:
The grism (gratings-prism) was proposed by Edwin Carpenter back in 1963.
A few words here in my official capacity as the developer of the Star Analyser. The use of a wedge prism with a grating used in the converging beam is not just to bring the spectrum onto the optical axis to minimise field curvature but to minimise all aberrations caused by the converging configuration.  The wedge prism supplied by Paton Hawksley is designed to do this for the SA100 when used as recommended by the manufacturer. These aberrations are small when the Star Analyser is used as recommended by the manufacturer. This is what the ,manufacturers says 
"
https://www.patonhawksley.com/product-page/3-8-prism
You can see the typical improvement in tests made by Christian Buil here

If using lower focal ratios and particularly with high dispersions or working in the IR, (ie outsed the normal operating range of the Star Analyser)  these aberrations become more apparent so adding a wedge prism can be particularly useful when designing high dispersion systems incorporating the Star Analyser.
 
--
www.threehillsobservatory.co.uk
Celestron C11, EQ6, ATIK 314, 428
Star Analyser 100,200, ALPY 600,200 LHIRES III Spectrographs


R L
 

OOPS pressed send before I finished !


On Fri, Aug 28, 2020 at 01:27 PM, R L wrote:
A few words here in my official capacity as the developer of the Star Analyser. The use of a wedge prism with a grating used in the converging beam is not just to bring the spectrum onto the optical axis to minimise field curvature but to minimise all aberrations caused by the converging configuration.  The wedge prism supplied by Paton Hawksley is designed to do this for the SA100 when used as recommended by the manufacturer. These aberrations are small when the Star Analyser is used as recommended by the manufacturer. This is what the ,manufacturers says 
"Designed for the advanced user looking for the maximum resolution, this device centralises the spectrum on the optical axis and reduces the small amount of coma sometimes visible in the red region of spectra captured with the Star Analyser 100.It is mounted in a 1.25" filter holder for direct attachment to the Star Analyser 100 and has been AR coated on both sides.Alignment markings indicate correct orientation with the Star Analyser 100. It can be mounted either behind the SA100 (adding 15mm to the distance to the camera sensor) and held in place by an optional locking ring, or ahead of it.Note that with the prism in place, the wavelength scale becomes significantly non linear so a non linear wavelength calibration should be used."
https://www.patonhawksley.com/product-page/3-8-prism

You can see an example of the typical improvement in tests made by Christian Buil when used in a typical SA100 setup  here
http://www.astrosurf.com/buil/staranalyser3/userguide.htm
third the way down the page 
"Improvised" mounting of a prism in front of the Star Analyzer. The prism is BK7 with an angle of 3.8°."
 With higher than recommended dispersions or working in the IR, (ie outside the normal operating range of the Star Analyser)  these aberrations become more apparent so adding a wedge prism can be particularly useful when designing high dispersion systems incorporating the Star Analyser.
Robin
 
--
www.threehillsobservatory.co.uk
Celestron C11, EQ6, ATIK 314, 428
Star Analyser 100,200, ALPY 600,200 LHIRES III Spectrographs"


R L
 

And Tim's observations and advice are spot on. Thanks Tim!

Cheers
Robin
--
www.threehillsobservatory.co.uk
Celestron C11, EQ6, ATIK 314, 428
Star Analyser 100,200, ALPY 600,200 LHIRES III Spectrographs


Tim Stone
 

And of course I misspelled "Aberrations" in the graphic... sigh


Christian Grainger <cgrainger2100@...>
 

Lol.



Sent from my Verizon, Samsung Galaxy smartphone

-------- Original message --------
From: Tim Stone <tim.stone.piano@...>
Date: 8/28/20 12:07 PM (GMT-05:00)
To: RSpec-Astronomy@groups.io
Subject: Re: [RSpec-Astronomy] Star Analyzer Prism Myths

And of course I misspelled "Aberrations" in the graphic... sigh


David Britz
 

Hi Tim,
I recently had an article published in Sky and Tel on my shoe box spectrograph. In essence the article described using an inch long polished stainless rod 1/4 inch dia and the Edmunds grating film all mounted in a foamcore box, with a camera located at the films exiting spectra and perpendicular to dispersion angle. Unfortunately S&T significantly reduced the article's coverage and never published the attached solar spectrum I acquired using a Canon 5D M2 and 100 mm lens. See attached PDF of the spectrum and spectrogram set up. It's clear even with the limited performance of the grating film the results are good!. I should try again with a proper transmission grating. The polished rod in effect produces a reflective solar glint "line" the length of the Sun's 1/2 degree lengthened by the camera's lens focal length, creating a very thin virtual reflective slit equivalent.  This was a single frame among many that I wanted to stack to bring out the lines (my previous threads subject on registax stacking). 
So yes polished rods do work very well, are cheap, the cylindrical nature of the refletor means the reflection slt edges stay perfectly aligned and give a "whole object" reflected spectra Vs the traditional slits small sampling of an extended objects focused  image on the slit - 
But there is also a faint but resolved "weird" periodic background structure visible in the green /red region that I assume is possibly a optical harmonic between the grating line resolution and the glint which i calculated to be less than a 10 micron width. Any thoughts anyone? 
The reflecting rod approach might be useful for in-line planetary spectra and imaging. From my experiments I believe an aluminum coated optically polished glass rod would be the best reflector, I have found the rod reflection approach is very sensitive to any microscopic surface imperfections on the rod, (the horizontal lines in the spectra). even though I used micro diamond polishing film to smooth and polish the rods surface in a lathe.

My casual visual inspection of the shoe box spectra revealed many high contrast but fine absorption lines that frustratingly I was not able to fully capture with the camera.Perhaps a video mode and some stacking may improve the results? 
I am not fully versed with RSpec yet, especially without a zero order reference source I am not sure how to calibrate and graph the spectrum. If someone in the group has a spare moment to do so I would be eternally grateful!

But I promise you, that after making one of these simple spectroscopes you will never look at a car's chrome trim reflecting sunlight the same way ever again. I can send the more extensive unpublished article if any one wants to take a look. 
Regards to all,
DaveB


When I decided to make the jump, I had to decide what hardware I wanted to use. I didn't have a grating, and so I needed to buy one. I bought the SA200, under the basic notion "If 100 is good, 200 is better." This not entirely true. It's similar to saying if a little screwdriver is good, a big old honking screwdriver is better. Like the screwdrivers, the SA100 and 200 are both good tools, and useful for what they are good for. It turns out, the 100 would have been a better learning grating for me. If you're just venturing into this space, I would recommend starting with the 100, and then adding the 200 if or when you decide you want to do something that tool is better for. But that's not what this post is about.



patrick kavanagh
 

Hello David:

Please forward your full article of your shoebox Solar Spectra imager and the stainless steel rod: your project sounds super!

Regards,

Patrick


On Sat, Aug 29, 2020 at 14:24, David Britz
<briswold@...> wrote:
Hi Tim,
I recently had an article published in Sky and Tel on my shoe box spectrograph. In essence the article described using an inch long polished stainless rod 1/4 inch dia and the Edmunds grating film all mounted in a foamcore box, with a camera located at the films exiting spectra and perpendicular to dispersion angle. Unfortunately S&T significantly reduced the article's coverage and never published the attached solar spectrum I acquired using a Canon 5D M2 and 100 mm lens. See attached PDF of the spectrum and spectrogram set up. It's clear even with the limited performance of the grating film the results are good!. I should try again with a proper transmission grating. The polished rod in effect produces a reflective solar glint "line" the length of the Sun's 1/2 degree lengthened by the camera's lens focal length, creating a very thin virtual reflective slit equivalent.  This was a single frame among many that I wanted to stack to bring out the lines (my previous threads subject on registax stacking). 
So yes polished rods do work very well, are cheap, the cylindrical nature of the refletor means the reflection slt edges stay perfectly aligned and give a "whole object" reflected spectra Vs the traditional slits small sampling of an extended objects focused  image on the slit - 
But there is also a faint but resolved "weird" periodic background structure visible in the green /red region that I assume is possibly a optical harmonic between the grating line resolution and the glint which i calculated to be less than a 10 micron width. Any thoughts anyone? 
The reflecting rod approach might be useful for in-line planetary spectra and imaging. From my experiments I believe an aluminum coated optically polished glass rod would be the best reflector, I have found the rod reflection approach is very sensitive to any microscopic surface imperfections on the rod, (the horizontal lines in the spectra). even though I used micro diamond polishing film to smooth and polish the rods surface in a lathe.

My casual visual inspection of the shoe box spectra revealed many high contrast but fine absorption lines that frustratingly I was not able to fully capture with the camera.Perhaps a video mode and some stacking may improve the results? 
I am not fully versed with RSpec yet, especially without a zero order reference source I am not sure how to calibrate and graph the spectrum. If someone in the group has a spare moment to do so I would be eternally grateful!

But I promise you, that after making one of these simple spectroscopes you will never look at a car's chrome trim reflecting sunlight the same way ever again. I can send the more extensive unpublished article if any one wants to take a look. 
Regards to all,
DaveB

When I decided to make the jump, I had to decide what hardware I wanted to use. I didn't have a grating, and so I needed to buy one. I bought the SA200, under the basic notion "If 100 is good, 200 is better." This not entirely true. It's similar to saying if a little screwdriver is good, a big old honking screwdriver is better. Like the screwdrivers, the SA100 and 200 are both good tools, and useful for what they are good for. It turns out, the 100 would have been a better learning grating for me. If you're just venturing into this space, I would recommend starting with the 100, and then adding the 200 if or when you decide you want to do something that tool is better for. But that's not what this post is about.



patrick kavanagh
 



Hello Tim:

I always enjoy reading your comments.

I just started this summer with my personal transfer from Astrophotography to Astronomical Spectroscopy, and I am still pretty much still at the starting line, but I realize that focusing, exposure and calibration in Spectroscopy is like learning a new language. 

I just copied your 11 point suggestions-for-beginners in the next-to-last paragraph, printed it and added it as an attachment to Tom's:Getting Started in Astronomical Spectroscopy with RSpec" Manual.

Thanks Tim.

Patrick


On Thu, Aug 27, 2020 at 21:13, Tim Stone
<tim.stone.piano@...> wrote:
There seems to be a significant uptick lately of interest in the Star Analyzer Prism. This prism is an accessory, designed specifically for use with the SA100 grating. Because I had some conceptions of the purpose of this prism that turned out to be simply and utterly incorrect, I thought I might relate those conceptions and hopefully clarify what the prism is actually for.

When I decided to make the jump, I had to decide what hardware I wanted to use. I didn't have a grating, and so I needed to buy one. I bought the SA200, under the basic notion "If 100 is good, 200 is better." This not entirely true. It's similar to saying if a little screwdriver is good, a big old honking screwdriver is better. Like the screwdrivers, the SA100 and 200 are both good tools, and useful for what they are good for. It turns out, the 100 would have been a better learning grating for me. If you're just venturing into this space, I would recommend starting with the 100, and then adding the 200 if or when you decide you want to do something that tool is better for. But that's not what this post is about.

I also bought the 3.8 degree prism. I thought it magically increased the resolution I would attain with my SA200. I was giddy with the thought that I could simply screw a wedge-shaped piece of glass onto my grating, and get even -better- results. At first, I acquired some spectra with my SA200 without the prism. I can tell you I was so thrilled to see absorption features, I could hardly contain myself. Imagine my thrill at the thought of doing EVEN BETTER than that with my prism. I screwed it on, acquired a spectrum from one of my previous targets, and... I didn't see any difference at all.

I thought I'd been ripped off. The thing clearly didn't do what I thought it would. Why did I spend all those $$ for it? In my disillusionment, I put it away and just concentrated on using the grating. That was the first correct thing I did.

Now, a couple of years later, I have come to understand the purpose of the prism, what it does, and what it does not do. Let's start there: what does it not do?

  1. It does not significantly improve your dispersion or resolution.
  2. It does not necessarily make your spectra better. I say -necessarily- because it -can- help in one way, I'll talk about in my list of what it -does- do.
  3. It is no substitute for quality data acquisition (e.g. focus, exposure), processing (e.g. stacking), and reduction (e.g. calibration) skills.
Let me dwell on point #3 for a minute. To even -see- the effect of the prism, you must be able to attain good focus. Focus is the #1 problem I had with my spectra. Even when I thought for sure I was in focus, I was way out of focus. I was chronically out of focus, no matter how little or much spacing I had between my prism and my sensor. If you don't know how to get excellent focus, then that's the only thing you should be working to perfect. Everything depends on that. (Hint: when the spectrum strip is in focus, the zero order star is not.)

Now, what -does- the prism do?

  1. It makes the spectrum strip distinctly non-linear. This is how prisms work. This introduces a lot of bother in the process of calibrating your spectra. You can no longer calibrate on the zero order and a well defined line. It won't work that way anymore. You will need to do at least a three-point calibration. If you're not really good at calibration, stay away from the prism. Once you can get excellent focus, then concentrate on becoming great at calibrating on two points. Graduate from needing the zero order as one of your points, to using two lines, one at the bluer end and and one at the redder end, to do your calibrations. Get good at this.
  2. For the bother (should I say "pain?") of non-linear calibration, you get a reduction in aberrations at a specific location in your spectrum image. The prism moves the spectrum strip toward the center axis of the optical system, where theoretically, aberrations such as coma and chromatic are null. The wavelength where the aberrations are null is determined by the grating and the prism deviation angle. For the SA 3.8 degree prism, the deviation angle is about two degrees. With the SA100 and the prism, at the grating/sensor spacing they were designed for, the aberration null wavelength should be about 6500A. With my SA200 and the prism at the spacing I use, the null is around 4300A. 



Once you've learned to acquire excellent data, process it well, and reduce (calibrate, normalize, etc.) it accurately, perhaps it is time to add a prism to your mix. It seems the null wavelength shifts with your grating/sensor spacing, so you will have to experiment some to position the wavelength you want to study at the null point. While you will not really have any higher resolution than without the prism, you will have lower aberration around the null point, and so your results there will likely be closer to the maximum for your system. The difference, though, is not nearly as dramatic as you might think.

Having been deeply involved in deep-sky astrophotography for years, I had gotten to the point where little things mattered to me. I was fine-tuning my deep-sky images in ways only I could see. When I stepped into spectrography, I had forgotten what it was like to be way down low on the learning curve, and naturally thought I understood everything well enough to continue to care about the little nitpicky things. It took me most of the first year to finally come to the realization that I didn't know anything about what I was doing, and needed to just go back to the fundamentals everyone here on this forum was trying to tell me about:

  • Practice, practice, practice
  • Learn to get awesome focus
  • Learn to get exposure correct
  • Learn to calibrate
  • Stick with bright stars
  • Learn with A stars first. 
  • Learn the cooler stars after the A stars.
  • Start tuning your spacing to see if you can get better results
  • Do some awesome stars like P-Cyg, a little dimmer, but still reachable
  • Do some Be stars or Mira variables
  • Now, is coma bothering you? Add a prism.

Last, let me offer my apology if this seems patronizing or pompous.I don't intend any offense. I don't know it all. I'm still only two years into this avocation. I may have some details wrong in this post. If so, I hope the forum will correct me! But... I'd just hate it if people had the same misconceptions I had about the prism and, like me, made decisions based on those misconceptions. If you're new, just know, there is a place for the prism, but you don't need it to take fantastic spectra!.


Tim Stone
 

Wow that's great, Patrick! I"m happy to help you in your journey any way I can :)


Tim Stone
 
Edited

On Sat, Aug 29, 2020 at 02:24 PM, David Britz wrote:
I am not sure how to calibrate and graph the spectrum. If someone in the group has a spare moment to do so I would be eternally grateful!
Here you go :) A few observations for you:

You've captured a good number of the Fraunhofer lines, and especially toward the blue end, quite a few other lines. Good job!

The system has some pretty severe astigmatism, which results in the out of focus condition on both ends. I can't tell if chromatic aberration is involved, but that might be exacerbating the focus, particularly on the blue end.

The oscillating pattern redward of about 5500A is not real. It might be from the grating film you used, or it might be the result of thin film interference on the sensor. This has been discussed fairly extensively on this forum, it's a characteristic of some sensors, and there's not much to be done about it other than use a different sensor. In my solar spectra, I get a similar pattern, and I'm using a high quality reflection grating. Just something to consider.

In a situation like this, there really isn't a zero order, which complicates calibration. Once you're familiar with the solar spectrum, the landmarks stick out like a sore thumb. I've labeled the ones present in your spectrum.

The spectrum is non-linear, which is why some of the lines appear slightly off from their reference wavelengths. I calibrated pixel 263 to 4308A and 2296 to 6563A, for your reference.

You don't have to be grateful for eternity ;)


David Britz
 

Hi Tim,
Thank you so much! Can I be grateful for at least a few days? 
I had been thinking of putting a collimating lens or cylindrical lens in-line between the rod reflector and the grating to create a collimated beam at least along the grating axis, which would help flatten the image plane and reduce aberrations,  the collimated beam going through the grating should also help with the nonlinearity. If the lens is large enough to illuminate the entire grating this should maximize the R value as well since a large number of grating lines would intercept the light. The false lines worry me, but you're right I do not recall seeing them when looking through the camera's optical viewfinder. Are the result you see here worth pursuing the slitlless concept? I am too much of a newby to judge that, and there are certainly a lot of things I can tweak. But if the cameras sensor is causing the false lines is there a video camera or DSLR  that you know does not suffer from the same issues.
Great work!
Regards,
DaveB

On Sat, Aug 29, 2020 at 7:38 PM Tim Stone <tim.stone.piano@...> wrote:
On Sat, Aug 29, 2020 at 02:24 PM, David Britz wrote:
I am not sure how to calibrate and graph the spectrum. If someone in the group has a spare moment to do so I would be eternally grateful!
Here you go :) A few observations for you:

You've captured a good number of the Fraunhofer lines, and especially toward the blue end, quite a few other lines. Good job!

The system has some pretty severe astigmatism, which results in the out of focus condition on both ends. I can't tell if chromatic aberration is involved, but that might be exacerbating the focus, particularly on the blue end.

The oscillating pattern redward of about 5500A is not real. It's (probably) the result of thin film interference on the sensor. This has been discussed fairly extensively on this forum, it's a characteristic of some sensors, and there's not much to be done about it other than use a different sensor.

In a situation like this, there really isn't a zero order, which complicates calibration. Once you're familiar with the solar spectrum, the landmarks stick out like a sore thumb. I've labeled the ones present in your spectrum.

The spectrum is non-linear, which is why some of the lines appear slightly off from their reference wavelengths. I calibrated pixel 263 to 4308A and 2296 to 6563A, for your reference.

You don't have to be grateful for eternity ;)


Tim Stone
 

Well as for whether you should work with this design to improve it, all I can say is if you want to do it, then do it :) I will tell you this, slitless solar spectra are neat as far as I'm concerned! One thing it might be useful for is outreach. Seeing the solar spectrum without having to point a telescope at the sun is useful. For that purpose, you might replace the camera with an eyepiece. If that worked, it might might give a really nice view. 

This has been done with a polished needle, which might give a narrower virtual slit... Tom shows one of those results on his projects page https://www.rspec-astro.com/sample-projects/

Also, yeah, I would think a collimator would help. An achromat might introduce some chromatic aberration, but if you stay at f10 or higher, that might not be too much of a factor. 

I can't answer the question about which sensors do not have this characteristic...


Jacques Savard
 

THE BRASS  WIRE  GIVE  THE BEST  RESULT

 

Jack 47 71

 

Well as for whether you should work with this design to improve it, all I can say is if you want to do it, then do it :) I will tell you this, slitless solar spectra are neat as far as I'm concerned! One thing it might be useful for is outreach. Seeing the solar spectrum without having to point a telescope at the sun is useful. For that purpose, you might replace the camera with an eyepiece. If that worked, it might might give a really nice view. 

This has been done with a polished needle, which might give a narrower virtual slit... Tom shows one of those results on his projects page https://www.rspec-astro.com/sample-projects/

Also, yeah, I would think a collimator would help. An achromat might introduce some chromatic aberration, but if you stay at f10 or higher, that might not be too much of a factor. 

I can't answer the question about which sensors do not have this character ic...


David Britz
 

Hi Jacques,
Thank you so much for your input and thoughts
Perhaps a low cost  "parabolic" condenser lens would solve the chromatic issues since it doesn't have to be a really great figure?
Darn it! The ancients have stolen my secrets once again! I had not seen the image below, so once again I am humbled by those who came before me.
Re the collimating lens I am thinking of a couple of different arrangements please see concept sketches, I would very much appreciate any thoughts on these layouts
Regards 
DaveB

image.png



On Sun, Aug 30, 2020 at 7:25 AM Jacques Savard <jacquessavard@...> wrote:

THE BRASS  WIRE  GIVE  THE BEST  RESULT

 

Jack 47 71

 

Well as for whether you should work with this design to improve it, all I can say is if you want to do it, then do it :) I will tell you this, slitless solar spectra are neat as far as I'm concerned! One thing it might be useful for is outreach. Seeing the solar spectrum without having to point a telescope at the sun is useful. For that purpose, you might replace the camera with an eyepiece. If that worked, it might might give a really nice view. 

This has been done with a polished needle, which might give a narrower virtual slit... Tom shows one of those results on his projects page https://www.rspec-astro.com/sample-projects/

Also, yeah, I would think a collimator would help. An achromat might introduce some chromatic aberration, but if you stay at f10 or higher, that might not be too much of a factor. 

I can't answer the question about which sensors do not have this character ic...


Ken Harrison
 

You illustration comes from Dave Howard's great webpage:
http://www.stargazing.net/david/spectroscopy/SimpleNeedleSpectroscope.html

I think you're over-thinking the issue. There's plenty of sunlight, and the glitter spot/ line is very very narrow. If positioned say >2 mtr from your 25mm aperture grating it's effectively at >f80 - a collimated beam.

--
"Astronomical Spectroscopy - The Final Frontier" - to boldly go where few amateurs have gone before....


R L
 

On Sun, Aug 30, 2020 at 05:20 PM, David Britz wrote:
I would very much appreciate any thoughts on these layouts
Regards 
The late Maurice Gavin did some experiments designing spectrographs with this type of setup using a shiny needle in place of a slit many years ago. His website has gone but it could be in the BAA journal somewhere. I will try to dig it out. Unlike the simple setup collecting sunlight from a needle at a distance, you definitely do need a collimating  lens in the design as you show if you want to use it on a telescope.  In fact this is the big drawback of this type of system. The problem is the glint from the needle is very, very low focal ratio, effectively the light goes in all directions so only a small fraction of the total light from the telescope goes through the collimator and into the spectrograph. This means it is very inefficient. It did work on bright stars, though. 

Robin
 
--
www.threehillsobservatory.co.uk
Celestron C11, EQ6, ATIK 314, 428
Star Analyser 100,200, ALPY 600,200 LHIRES III Spectrographs


R L
 

On Mon, Aug 31, 2020 at 03:25 PM, R L wrote:
The problem is the glint from the needle is very, very low focal ratio, effectively the light goes in all directions so only a small fraction of the total light from the telescope goes through the collimator and into the spectrograph.
The last, afocal design on your pdf where you collimate the light before the needle would be doubly inefficient as most of the light from the star would miss the needle. The needle (or slit or fibre etc) needs to be at the focal plane of the telescope as you show in the version above this one.

Cheers
Robin
 
--
www.threehillsobservatory.co.uk
Celestron C11, EQ6, ATIK 314, 428
Star Analyser 100,200, ALPY 600,200 LHIRES III Spectrographs