56 replies to this topic

[eavaria]

Hello. I'm thinking in long term projects, and the fllowing idea came to my mind...

 

Would it be possible to get a singlet refractor and use beam splitters to get every range of frequencies into a different monochromatic sensor at the side of the main beam, being able to focus perfectly every color despise tbe CA of the lens itself.

 

This will be more expensive and complex than an APO of the same characteristics, but once the system is built, it will be really y easy (and cheap) to replace the main lens and have big refractors for every purpose just changimg the singlet.

 

Anyone have seen something like that, or have any thought on this? I believe that the space for the splitters won't be big enough for relatively short focal lenghts, so probably another lens(es) need to be added to get more space for the splitters, buy I believe it's doable... I can see myself having an F3000+ and an F600 refractor in the same telescope using cheap high diameter singlets... 

[LDW47]

Hello. I'm thinking in long term projects, and the fllowing idea came to my mind...

 

Would it be possible to get a singlet refractor and use beam splitters to get every range of frequencies into a different monochromatic sensor at the side of the main beam, being able to focus perfectly every color despise tbe CA of the lens itself.

 

This will be more expensive and complex than an APO of the same characteristics, but once the system is built, it will be really y easy (and cheap) to replace the main lens and have big refractors for every purpose just changimg the singlet.

 

Anyone have seen something like that, or have any thought on this? I believe that the space for the splitters won't be big enough for relatively short focal lenghts, so probably another lens(es) need to be added to get more space for the splitters, buy I believe it's doable... I can see myself having an F3000+ and an F600 refractor in the same telescope using cheap high diameter singlets... 

You have to be kidding ? Right ? All for a simple nite of viewing ?

[eavaria]

Not sure what you mean. Actually it's more like a theoretical question. The idea is to have a complex data capture device that can deal with the optical issues that have been there for decades. For the same reason, I'm pretty sure I'm missing something that makes this impossible...

[LDW47]

Not sure what you mean. Actually it's more like a theoretical question. The idea is to have a complex data capture device that can deal with the optical issues that have been there for decades. For the same reason, I'm pretty sure I'm missing something that makes this impossible...

Its your post, clarify, explain more, who are you directing your questions to ? More expanded details please !

[sg80]

I am not the sharpest knife in the drawer. but if there was something like a digital light converter at the back end of the scope. that collected the light and put red green and blue together right before the eyepiece.

[LDW47]

I am not the sharpest knife in the drawer. but if there was something like a digital light converter at the back end of the scope. that collected the light and put red green and blue together right before the eyepiece.

One more time ‘ you have to be kidding ‘ ?

[dan_h]

I am not the sharpest knife in the drawer. but if there was something like a digital light converter at the back end of the scope. that collected the light and put red green and blue together right before the eyepiece.

Isn't this basically what a color camera does? 

 

dan

[sg80]

One more time ‘ you have to be kidding ‘ ?LDW47       No I'm not what I am asking do you think it's possible to use a digital fix for chromatic aberration. I do some home recording and I have a program that does music re balance it can hep fix mixed recordings that are not right bring up the vocals or drums or bass.so lets say you can make a digital diagonal the light goes from the object lens to the diagonal a processor  balances the light and fixes the chromatic aberration ahead of the eyepiece. analog to digital then digital to analog converters.a visual re balance just input the lens diameter 102mm  f10 achro and let it go. I have to admit the music re balance does not work all the but it does work most of the time and the software was not real expensive. this if possible would be hardware but could be updated.i do not want to speak for the original poster but i think he just wanted to know if chromatic aberration could be fixed in a different way.he could be on to something or maybe not the point here is this hobby is dwindling in numbers.we should welcome interest in this hobby.and our responses should be to help or educate.

Edited by sg80, 05 December 2020 - 09:37 PM.

[sg80]

Hello. I'm thinking in long term projects, and the fllowing idea came to my mind...

 

Would it be possible to get a singlet refractor and use beam splitters to get every range of frequencies into a different monochromatic sensor at the side of the main beam, being able to focus perfectly every color despise tbe CA of the lens itself.

 

This will be more expensive and complex than an APO of the same characteristics, but once the system is built, it will be really y easy (and cheap) to replace the main lens and have big refractors for every purpose just changimg the singlet.

 

Anyone have seen something like that, or have any thought on this? I believe that the space for the splitters won't be big enough for relatively short focal lenghts, so probably another lens(es) need to be added to get more space for the splitters, buy I believe it's doable... I can see myself having an F3000+ and an F600 refractor in the same telescope using cheap high diameter singlets... 

I think you are thinking outside the box great

[LDW47]

One more time ‘ you have to be kidding ‘ ?LDW47       No I'm not what I am asking do you think it's possible to use a digital fix for chromatic aberration. I do some home recording and I have a program that does music re balance it can hep fix mixed recordings that are not right bring up the vocals or drums or bass.so lets say you can make a digital diagonal the light goes from the object lens to the diagonal a processor  balances the light and fixes the chromatic aberration ahead of the eyepiece. analog to digital then digital to analog converters.a visual re balance just input the lens diameter 102mm  f10 achro and let it go. I have to admit the music re balance does not work all the but it does work most of the time and the software was not real expensive. this if possible would be hardware but could be updated.i do not want to speak for the original poster but i think he just wanted to know if chromatic aberration could be fixed in a different way.he could be on to something or maybe not the point here is this hobby is dwindling in numbers.we should welcome interest in this hobby.and our responses should be to help or educate.

I guess its because I never found CA a big problem with any of my refractors (10) that range f4.8-f15, just lucky I guess or just not picky, lol ! So I really don’t need ............. ! Carry on, nothing wrong with reinventing the wheel, if its needed by some ?

Edited by LDW47, 05 December 2020 - 10:09 PM.

[sg80]

I guess its because I never found CA a big problem with any of my refractors (10) that range f4.8-f15, just lucky I guess or just not picky, lol ! So I really don’t need ............. ! Carry on, nothing wrong with reinventing the wheel, if its needed by some ?

It may not be possible it may and it could low cost or more expensive.just throwing ideas out there.also at first I did not notice CA until it was talked on this forum years ago.after reading the thread I looked through my scope and there it was Thanks Guys lol but it is not to bad.

[sg6]

Dicroic filters alone could not work, the order they would have to sit prevents it.

Light would have to be split via beam splitters then each split beam fed into a filter system. Beam splitters would halve the light at each split by say 50%. So you have differing "intensities". First would be 50% then 2 at 25%, that is if you had a 3 way split - the last is not split just reflected off a mirror. So you get 50%, 25%, 25%.

 

Have to remember that a doublet will do something for chromatic aberration but also hopefully do a little for spherical aberration. Which a singlet in a way cannot. Best on a singlet is radaii in the ratio (approximately) of 1:6 for SA.

 

Could likely be set up on an optical bench in a lab, but quality of at the end would be seriously questionable, don't expect anything sharp. Every bit of glass added will degrade the end result to some extent. Even to this day a simple well made plossl is still a good eyepiece. A fair % of the glass in multi-element eyepieces is there to correct for problems caused by other bits of glass.

 

Start at 4 element plossl, to get a bit wider add in 5th element, that causes soft edges/curvature so add a 6th in to correct for that, now it is 6 elements. Do you need another for increased CA? And so it goes on.

 

Also the idea does not fix CA, each wavelength will still be at a different focal plane all you have done is split the light and then imaged each on to a sensor at different  focal planes. The CA is still present as the although separate out the focal lengths are different. Which is CA: Different colors focus at different focal planes.

Edited by sg6, 06 December 2020 - 05:08 AM.

[eavaria]

For context, I'm a software engineer, and when thinking of fixing CA using software, I came to the conclusion that it was impossible because the aberration is light that is actually out of focus, so infirmation is lost when photons hit the sensor (the direction is partially lost). So to be able to fix it, we need it to be focused, and that's when more optical surfaces come into play...
The thing is the challenges of triplets and more are still a thing, I was thinking that if the beam can be splitted and frequencies focused and processed by themselves, we could use high quality glass cheap doublets with tbis reusable complex part, makimg it convenient in tje case of a. 300mm doublet tvat is payable, unlike a 300mm APO... In tje same page, the same beam splitter could be used with a 400mm and a 3000mm, potentially lowering the total cost (400mm APO + 3000mm APO vs 400mm doublet + 3000mm doublet + splitter) and being able to push the current limits... I'm just trying to find if it's worth making the experiment myself with some cheap chinese doublets, filters and ccds... Apparently it does 😊

[eavaria]

For context, I'm a software engineer, and when thinking of fixing CA using software, I came to the conclusion that it was impossible because the aberration is light that is actually out of focus, so infirmation is lost when photons hit the sensor (the direction is partially lost). So to be able to fix it, we need it to be focused, and that's when more optical surfaces come into play...
The thing is the challenges of triplets and more are still a thing, I was thinking that if the beam can be splitted and frequencies focused and processed by themselves, we could use high quality glass cheap doublets with tbis reusable complex part, makimg it convenient in tje case of a. 300mm doublet tvat is payable, unlike a 300mm APO... In tje same page, the same beam splitter could be used with a 400mm and a 3000mm, potentially lowering the total cost (400mm APO + 3000mm APO vs 400mm doublet + 3000mm doublet + splitter) and being able to push the current limits... I'm just trying to find if it's worth making the experiment myself with some cheap chinese doublets, filters and ccds... Apparently it does 😊

[sg80]

For context, I'm a software engineer, and when thinking of fixing CA using software, I came to the conclusion that it was impossible because the aberration is light that is actually out of focus, so infirmation is lost when photons hit the sensor (the direction is partially lost). So to be able to fix it, we need it to be focused, and that's when more optical surfaces come into play...
The thing is the challenges of triplets and more are still a thing, I was thinking that if the beam can be splitted and frequencies focused and processed by themselves, we could use high quality glass cheap doublets with tbis reusable complex part, makimg it convenient in tje case of a. 300mm doublet tvat is payable, unlike a 300mm APO... In tje same page, the same beam splitter could be used with a 400mm and a 3000mm, potentially lowering the total cost (400mm APO + 3000mm APO vs 400mm doublet + 3000mm doublet + splitter) and being able to push the current limits... I'm just trying to find if it's worth making the experiment myself with some cheap chinese doublets, filters and ccds... Apparently it does

I was talking about using a doublet.then a sensor that collects the light then software puts the info in focus or approximate the image for human vision something like video or audio upscaling. I not a engineer of any type but I am impressed by some of the things that has been done for audio by companies such as Izotope Sony and others just wondering if it could be used for visual.

[gezak22]

For context, I'm a software engineer, and when thinking of fixing CA using software, I came to the conclusion that it was impossible because the aberration is light that is actually out of focus, so infirmation is lost when photons hit the sensor (the direction is partially lost). So to be able to fix it, we need it to be focused, and that's when more optical surfaces come into play...
The thing is the challenges of triplets and more are still a thing, I was thinking that if the beam can be splitted and frequencies focused and processed by themselves, we could use high quality glass cheap doublets with tbis reusable complex part, makimg it convenient in tje case of a. 300mm doublet tvat is payable, unlike a 300mm APO... In tje same page, the same beam splitter could be used with a 400mm and a 3000mm, potentially lowering the total cost (400mm APO + 3000mm APO vs 400mm doublet + 3000mm doublet + splitter) and being able to push the current limits... I'm just trying to find if it's worth making the experiment myself with some cheap chinese doublets, filters and ccds... Apparently it does

The issue is that red (or green or blue) is not one wavelength. It's still a rather wide range of wavelengths (~100 nm) and a singlet will still produce plenty of CA over that range. So you'll have to split the light up into N different wavelengths, not just three. So if you want to collect 1 hour worth of information, you need to either collect N hours of data (1 hour for each of the N wavelenths), or you need N detectors. Note that this just solves CA, and a singlet will have other aberrations.

 

So, why make such an overly complex solution to a problem that can be solved with a single mirror?

 

Edit: If we are talking about using a doublet rather than an singlet, this becomes more feasible. Still, if CA is critical, why not use optics that are inherently free of CA?

Edited by gezak22, 06 December 2020 - 09:33 PM.

[RichA]

Hello. I'm thinking in long term projects, and the fllowing idea came to my mind...

 

Would it be possible to get a singlet refractor and use beam splitters to get every range of frequencies into a different monochromatic sensor at the side of the main beam, being able to focus perfectly every color despise tbe CA of the lens itself.

 

This will be more expensive and complex than an APO of the same characteristics, but once the system is built, it will be really y easy (and cheap) to replace the main lens and have big refractors for every purpose just changimg the singlet.

 

Anyone have seen something like that, or have any thought on this? I believe that the space for the splitters won't be big enough for relatively short focal lenghts, so probably another lens(es) need to be added to get more space for the splitters, buy I believe it's doable... I can see myself having an F3000+ and an F600 refractor in the same telescope using cheap high diameter singlets... 

There have been some solar scopes (HA) using singlet lenses because they only need a narrow wavelength to image.  But they've been long focal lengths to minimize aberrations like SA. You'd need very steep cut-off filters and despite having three sensors, you will sacrifice some of the incoming light, but who knows, maybe it will be more effective than dealing conventionally with CA?

[Wildetelescope]

Dicroic filters alone could not work, the order they would have to sit prevents it.

Light would have to be split via beam splitters then each split beam fed into a filter system. Beam splitters would halve the light at each split by say 50%. So you have differing "intensities". First would be 50% then 2 at 25%, that is if you had a 3 way split - the last is not split just reflected off a mirror. So you get 50%, 25%, 25%.

 

Have to remember that a doublet will do something for chromatic aberration but also hopefully do a little for spherical aberration. Which a singlet in a way cannot. Best on a singlet is radaii in the ratio (approximately) of 1:6 for SA.

 

Could likely be set up on an optical bench in a lab, but quality of at the end would be seriously questionable, don't expect anything sharp. Every bit of glass added will degrade the end result to some extent. Even to this day a simple well made plossl is still a good eyepiece. A fair % of the glass in multi-element eyepieces is there to correct for problems caused by other bits of glass.

 

Start at 4 element plossl, to get a bit wider add in 5th element, that causes soft edges/curvature so add a 6th in to correct for that, now it is 6 elements. Do you need another for increased CA? And so it goes on.

 

Also the idea does not fix CA, each wavelength will still be at a different focal plane all you have done is split the light and then imaged each on to a sensor at different  focal planes. The CA is still present as the although separate out the focal lengths are different. Which is CA: Different colors focus at different focal planes.

Yep!  This captures the implications of the physics nicely.   Hard to beat a well executed Doublet or Triplet design.  It very efficiently puts light where it is supposed to go.  

 

Here is another thought to consider.   What the OP is describing is Narrow band imaging.   And as we all know, there are a lot of talented folks making great pictures doing it.  This approach is ALSO somewhat more tolerant of objective lenses that fall short of Apochromatic perfection:-). However, as sg6 points out, those extra elements do more than just bring Red, Green and Blue into the same focal plane, they also help correct for other optical aberrations that are inherent in singlet lens.  There are no free lunches.

 

Cheers!

 

JMD

[Jon Isaacs]

For context, I'm a software engineer, and when thinking of fixing CA using software, I came to the conclusion that it was impossible because the aberration is light that is actually out of focus, so infirmation is lost when photons hit the sensor (the direction is partially lost). So to be able to fix it, we need it to be focused, and that's when more optical surfaces come into play...
The thing is the challenges of triplets and more are still a thing, I was thinking that if the beam can be splitted and frequencies focused and processed by themselves, we could use high quality glass cheap doublets with tbis reusable complex part, makimg it convenient in tje case of a. 300mm doublet tvat is payable, unlike a 300mm APO... In tje same page, the same beam splitter could be used with a 400mm and a 3000mm, potentially lowering the total cost (400mm APO + 3000mm APO vs 400mm doublet + 3000mm doublet + splitter) and being able to push the current limits... I'm just trying to find if it's worth making the experiment myself with some cheap chinese doublets, filters and ccds... Apparently it does

 

As I understand it, you propose to use beam splitters to separate the light into a number of individual colors and then sample them individually and recreate the images from the separate colors digitally. I see a number of issues that make this impractical.

 

- Each beam splitter reduces the transmission by a factor of two.  This represents a large loss of light.

 

- Each color will need to be aligned and focused individually.  Focus depends on temperature.  This is a nightmare, focusing just one optic is difficult as focus shifts throughout the night. It's hard to imagine focusing multitple sensors.  

 

- Refractors suffer from field curvature and the difference in focal length of the various colors (this is what chromatic aberration is), depends on the focal length of objective.  With each objective swap, you would have to realign and refocus in significant way. 

 

- Field curvature means that the edge of the field is not in focus when the center is in focus. This is a function of focal length.  Each time you swap objectives, you would need to swap the field flattener optics. This is tricky and time consuming. 

 

- Besides light, resolution would almost certainly be lost. 

 

- Inexpensive achromats suffer more than just chromatic aberration.. 

 

It seems like a very complex optical system that is compromised in a many ways.  besides the beam splitters, there will be numerous microstages and mirrors all mounted on an optical bench and these will be relatively large mirrors because they have to be larger than the focal plane.

 

Realize, all this optics has to be placed before the focal plane so you will be working with a relatively large beam that is not in focus.   

 

All this to avoid using better quality glass.. Or as Geza said, why not just use an mirror? a Newtonian is perfectly corrected for chromatic aberration.  

 

One rule of thumb in optics is to keep it as simple as possible, each element adds aberrations. Quality refractors are much better than diffraction limited.  That means they provide an image that is perfect to a small fraction of a wave length of light.  

 

Jon

[eavaria]

Singlets are out of the equation. Spherical aberration is still an issue and doublets are cheap enough. Triplets and more are the ones that become an issue.

I've read the comments, and dichroic mirrors are available in all ranges of wavelenghts. They can be manufactured at will, and magenta mirrors are readily available that reflects 95-99% of blue and red (depending on quality,they cost $5-$120) and transmit 90% of green, already mimicking what's currently being done with doublets with green treatment to filter out CA. The total power is split, but the power of the relevant light is kept 90%+

The other issues (focus changes, other aberrations and field flattening and the complexity of the optics is exactlt what I'm trying to test. I'm not saying that this will be practical. I'm considering it's possible... If it can be made practical is another story in the future... That's for example is what big data did about 10 years ago (I'm a big data engineer) ... Everything was done in a really complex way, but that allowed to overcome some hard barriers that were otherwise impossible to deal with... Then throwing more hardware into it solved the new issues (today a double cpu, 48 threads and 64 gb ram can be found for $400 or less...) that's what I believe can be done here... Every issue that doesn't involve losing information can be fixed by software... That includes field flattening, color shifts, vignetting, or field microfocusing... As long as you have the issue isolated enough to be able to deal with it without touching the rest of the image... That's when splitting by frequency comes into play...

I'm planning a test with a 72mm doublet at low focal lenght (less than 400) with just a magenta mirror to be able to test it against my Askar FRA400... The issue now is that all doublets come already with green or blue filtered out precisely to minimize CA... I'm trying to get a doublet without any screen.

The problem of needing multiple detectors is real, but I believe that there's a spot where the benefits outweights the cost of adding more detectors as adding more detectors grows linearly while lens size cost grows exponentially, and cheap specialized cmos should be a thing in the near future...

I plan to test with cheap 2MP detectors by now as a POC. Haven't decided yet how cheap, but hopefully in the range of $50 each... Of course will make sure that they are sensible enough... If anyone has a recommendation or comment on detectors, please speak up 😊

Thanks all for your input.

[EJN]

The below image of NGC 6819 was taken with a 85mm f/10 achromat using a #11 yellow filter. After stacking the RGB channels were realigned to give "natural" colors. An f/10 is pretty forgiving of CA. For a faster achromat, you could use a filter wheel with 3 narrowband filters. You would have to refocus when changing filters. No beamsplitters or multiple sensors required.

 

Sometimes it is better not to overthink a problem, it leads to a complex solution where a simpler solution already exists. Google "Rube Goldberg."

 

Click image for uncompressed view.

 

 

 

 

 

The issue now is that all doublets come already with green or blue filtered out precisely to minimize CA...

 

I think you are confused, doublets bring 2 colors to a common focus. The coatings are anti-reflection coatings, not filters. A few lenses were made with minus violet coatings applied, but that is a rare exception.

Edited by EJN, 07 December 2020 - 12:07 PM.

[Jon Isaacs]

GThe other issues (focus changes, other aberrations and field flattening and the complexity of the optics is exactlt what I'm trying to test. I'm not saying that this will be practical. I'm considering it's possible... If it can be made practical is another story in the future... That's for example is what big data did about 10 years ago (I'm a big data engineer) ... Everything was done in a really complex way, but that allowed to overcome some hard barriers that were otherwise impossible to deal with... Then throwing more hardware into it solved the new issues (today a double cpu, 48 threads and 64 gb ram can be found for $400 or less...) that's what I believe can be done here... Every issue that doesn't involve losing information can be fixed by software... That includes field flattening, color shifts, vignetting, or field microfocusing... As long as you have the issue isolated enough to be able to deal with it without touching the rest of the image... That's when splitting by frequency comes into play...

 

I worked in research in a laboratory and we did quite a bit of optics. It seems to me, you are minimizing the difficulties and expenses of this scheme.

 

Have you ever worked in a laboratory doing hands on optics?  This is not a computer, this is light that has not reached the focal plane.

 

Have you considered the requirements for the optical quality of the filters? How much do the filters cost of sufficient quality?  Amateur Astronomers commonly use thin film interference filters for enhancing the contrast of certain nebulae, a good 48 mm filter is about $200.

 

When you reflect light multiple times, the optical imperfections are additive. This means even higher quality surfaces are required.

 

Have you drawn up the optics with the appropriate angles for the reflections to the sensors and the effect on the light path and bandwidth of the filter? It seems like the filters will need to be tilted. 

 

Can this be accomplished relatively close to the focal plane so all the sensors can be at prime focus?

 

It's a lot work to solve a problem that can be easily solved with a parabolic mirror.  Chromatic aberration is no more.  

 

Jon

[LDW47]

I worked in research in a laboratory and we did quite a bit of optics. It seems to me, you are minimizing the difficulties and expenses of this scheme.

 

Have you ever worked in a laboratory doing hands on optics?  This is not a computer, this is light that has not reached the focal plane.

 

Have you considered the requirements for the optical quality of the filters? How much do the filters cost of sufficient quality?  Amateur Astronomers commonly use thin film interference filters for enhancing the contrast of certain nebulae, a good 48 mm filter is about $200.

 

When you reflect light multiple times, the optical imperfections are additive. This means even higher quality surfaces are required.

 

Have you drawn up the optics with the appropriate angles for the reflections to the sensors and the effect on the light path and bandwidth of the filter? It seems like the filters will need to be tilted. 

 

Can this be accomplished relatively close to the focal plane so all the sensors can be at prime focus?

 

It's a lot work to solve a problem that can be easily solved with a parabolic mirror.  Chromatic aberration is no more.  

 

Jon

Well said, I am not an expert by any means but you have to know it was not a simple matter at a cheap, affordable cost to the average astronomer, in the end ! And for what ? 

[dan_h]

 They can be manufactured at will, and magenta mirrors are readily available that reflects 95-99% of blue and red (depending on quality,they cost $5-$120) and transmit 90% of green, already mimicking what's currently being done with doublets with green treatment to filter out CA. 

...............I'm planning a test with a 72mm doublet at low focal lenght (less than 400) with just a magenta mirror to be able to test it against my Askar FRA400... The issue now is that all doublets come already with green or blue filtered out precisely to minimize CA...


Thanks all for your input.

 

It seems you are making some assumptions about lenses that are wrong at the most basic level. The idea that anti reflection coatings are in some way a filter applied to reduce CA of the lens is just incorrect.   These coatings are applied to increase the transmission of light and reduce ghosting as well as improve overall light grasp.   The reflected light you can see is a very small percentage of the overall light and in most cases will not impact the perceived colors in the final image. CA is controlled by the dispersion of the glass type(s) chosen for the lens.   

 

There are a number of excellent articles written on how to design an achromatic lens  You may find it helpful to work through a couple of these to improve your understanding of the problem you are trying to correct.  

 

dan

[eavaria]

 

I think you are confused, doublets bring 2 colors to a common focus. The coatings are anti-reflection coatings, not filters. A few lenses were made with minus violet coatings applied, but that is a rare exception.

At least chinese acromats come explicitly with a green screen between the flint and crown... They filter out the green light to reduce CA. That's separate from the coatings...