I agree with you that someone could do some test.
i guess that the setup with 10' throw is the one that needs 20mm between frresnell and lcd, is it ok? i guess the larger thows should have more than 20mm so you are focusing at center again.
is it what you mean?
i have done some math and found my OPTIMAL REAL WORLD (arc lengh and triplet 135 included design) would work perfect with 600mm focal field. You see it is very close from yours (550) as it can be adjusted with the lamp positioning, there is no problem at all.
Anyway there is still something i am working on, the triplet's max field angle... it is very important to be kown, if it is not sufficent, all the design is not correct. Do you agree?
i guess that the setup with 10' throw is the one that needs 20mm between frresnell and lcd, is it ok? i guess the larger thows should have more than 20mm so you are focusing at center again.
is it what you mean?
i have done some math and found my OPTIMAL REAL WORLD (arc lengh and triplet 135 included design) would work perfect with 600mm focal field. You see it is very close from yours (550) as it can be adjusted with the lamp positioning, there is no problem at all.
Anyway there is still something i am working on, the triplet's max field angle... it is very important to be kown, if it is not sufficent, all the design is not correct. Do you agree?
real testing
Yes, I agree that the maximum field angle is a very important parameter to know. If that is not wide enough for the LCD size and the LCD to projection lens distance, then it will not work very well.
I have read at least one post on this forum or DIYBG, that claimed to have a good even image using the 450 mm fl 135 mm diameter triplet with a 17" LCD. If that is true, it means that the lens maximum field angle is okay for 17" LCDs. It also suggests that dim corners are caused by problems with the fresnel focal lengths and positions, or bad light engines.
If I had one of those triplets, I would do some experiments and post the results (including the maximum field angle). Unfortunately, I don't have one.
Yes, I agree that the maximum field angle is a very important parameter to know. If that is not wide enough for the LCD size and the LCD to projection lens distance, then it will not work very well.
I have read at least one post on this forum or DIYBG, that claimed to have a good even image using the 450 mm fl 135 mm diameter triplet with a 17" LCD. If that is true, it means that the lens maximum field angle is okay for 17" LCDs. It also suggests that dim corners are caused by problems with the fresnel focal lengths and positions, or bad light engines.
If I had one of those triplets, I would do some experiments and post the results (including the maximum field angle). Unfortunately, I don't have one.
If anyone wants to send me a sample I’d be more than happy to test it. I just thought I'd save Guy the hassle. 
😀
DJ
😀 DJ
there is no problem working with 135 triplet and 17" lcd until you get to a large throw, i guess you knoe that. There is people saying he had succes with the triplet and the crt tv fast test. But then you ask them about the throw and you will see that the throw they used was very short (it does triplet --lcd distance relatively large so the 17" fit's inside FOV.
this is my friends first projections with 135 native and 17" with 10' throw corner close photo.
this is my friends first projections with 135 native and 17" with 10' throw corner close photo.
image
It does look like the corners are dim, but you can't really tell if it is a field angle problem just from a screen image.
Maybe the LCD is not lit very well at the corners by the fresnels.
Maybe the corners are too far from the center for the LCD's viewing angle. (Assuming a non-split design.)
Maybe most of the corner light is not getting through the lens because the the arc image is focussed to far from the optical center of the lens.
What are the fresnel focal lengths and lamp arc to condensor fresnel spacing of your friend's projector?
It does look like the corners are dim, but you can't really tell if it is a field angle problem just from a screen image.
Maybe the LCD is not lit very well at the corners by the fresnels.
Maybe the corners are too far from the center for the LCD's viewing angle. (Assuming a non-split design.)
Maybe most of the corner light is not getting through the lens because the the arc image is focussed to far from the optical center of the lens.
What are the fresnel focal lengths and lamp arc to condensor fresnel spacing of your friend's projector?
the fresnells are 330/330 unsplit.
the blurr image could be because of the inproper fresnell focal, but i would like to show you new pic, This one is with modified triplet (larger focal)
this one is better focused at corners than in center (so field fresnell impropper focal is not the reason of blurr. This is a quick test with the triplet, he has a better mod since then.
the light distribution is not the best, but all 17" can be seen, in a better way than the native 135 with; remenber 330/330.
this is one full screen quick test with mod triplet;
the blurr image could be because of the inproper fresnell focal, but i would like to show you new pic, This one is with modified triplet (larger focal)
An externally hosted image should be here but it was not working when we last tested it.
this one is better focused at corners than in center (so field fresnell impropper focal is not the reason of blurr. This is a quick test with the triplet, he has a better mod since then.
the light distribution is not the best, but all 17" can be seen, in a better way than the native 135 with; remenber 330/330.
this is one full screen quick test with mod triplet;
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
this one is the "best" full screen result i have seen with 135 native and 17" (doncuppjr at Diylabs). (the trans is somewhere 14" but the white square diagonal is 17" (16:9) see how the corners are black. we can´t say nothing about the focus on the corner because there is no image and if there was image, isn´t close enough to check it.
but diyeitors pic are very close, and are at corner. The only problem is the light uniformity with him. (the pics i have posted were with modified triplet and the lamp was not at the best position, he says it where i took the pics from.
those pics are not final mod pics.
Anyway he says that there is a great improvement since he adventured on striping the triplet. I´m still waiting final setup close up pics from him.
But if we compare results of doncuppirjd and diyeitor, i prefer diyeitor's. Don´t you?
Anyway he says that there is a great improvement since he adventured on striping the triplet. I´m still waiting final setup close up pics from him.
But if we compare results of doncuppirjd and diyeitor, i prefer diyeitor's. Don´t you?
Guy Grotke said "we see nothing positive and many negative effects from doing that", he reffers to filling rear lens aperture with light cone, thats what i would do and am triyng to xplain why.
Lets see a example; We have two projection systems; 1)difuse source and 2)Fresnell type source (not difuse but unidirectional light source)
this is diffuse
this is fresnell based.
Now think we block half of the light on the output just after the lens;
The resulting image will be dimmer (half lumens exactly) but the focus will be aparently better. (Why better? think that the other half will be hiperprjected on top, so litlle error will be between them that will become on less sharpness).
this would be the fresnell based projection with half light. But now, the image will be half black and half perfect (same brightness and same sharpness on the working half)
That's because each lens small area is working on each lcd esclusive area so the image is not being hiperdrawn one on top to the next, so the quality is best overall. And here i come to the point i was triyng to get; LARGER the lens, MORE EXCLUSIVE work will do each small lens area with the lcd correspondent area, so better projected image i guess.
I mean larger the lens is more area hitting by the light cone. So we should use 100% of our rear lens aperture, a thing that we will not do with 550mm field fresnell (1cm outer circle segment is not being used, i know is not to much but 600mm fresnell will have 0 unused area).
Lets see a example; We have two projection systems; 1)difuse source and 2)Fresnell type source (not difuse but unidirectional light source)
An externally hosted image should be here but it was not working when we last tested it.
this is diffuse
An externally hosted image should be here but it was not working when we last tested it.
this is fresnell based.
Now think we block half of the light on the output just after the lens;
An externally hosted image should be here but it was not working when we last tested it.
The resulting image will be dimmer (half lumens exactly) but the focus will be aparently better. (Why better? think that the other half will be hiperprjected on top, so litlle error will be between them that will become on less sharpness).
An externally hosted image should be here but it was not working when we last tested it.
this would be the fresnell based projection with half light. But now, the image will be half black and half perfect (same brightness and same sharpness on the working half)
That's because each lens small area is working on each lcd esclusive area so the image is not being hiperdrawn one on top to the next, so the quality is best overall. And here i come to the point i was triyng to get; LARGER the lens, MORE EXCLUSIVE work will do each small lens area with the lcd correspondent area, so better projected image i guess.
I mean larger the lens is more area hitting by the light cone. So we should use 100% of our rear lens aperture, a thing that we will not do with 550mm field fresnell (1cm outer circle segment is not being used, i know is not to much but 600mm fresnell will have 0 unused area).
all based on impossible design!
Rox, your drawings of the LCD-based designs would only be true if you had a very small point-source lamp. But we don't use $1000 lamps! We use $25 to $60 US lamps with 16 to 25 mm long arcs. So it just never occurs that "Each lens small area working on a specific LCD area".
If you remove the triplet and put a piece of white paper where the triplet goes, then you can see what I mean. Use your computer to generate an image for the LCD with a very small circle of white in the center surrounded by all black. You will not see a small circle of light on the paper! Because of the large lamp arc, you will see a much larger circle on the paper. Every part of the lamp arc will shine through a single pixel and spread that light over the entire size of the projected arc image.
If the lamp arc is 24 mm long and you use 220 and 550 mm fl fresnels, then you will get around a 60 mm arc image through every LCD pixel:
arc image size = 24 * (550 / 220)
If you use the right focal length fresnels, you can make all of those 60 mm circles fall at the center of the projection lens. If you use longer fresnels, then those 60 mm circles will each use an area from the center of the lens all the way to the outer diameter of the lens. Since the center area of a lens gives you a much higher image quality than the edges, longer focal length fresnels will give you an image that is not as sharp. If you use even longer fresnels, then some part of the 60 mm circles from the LCD corners will fall past the first lens surface, so the corners will be dim.
Rox, your drawings of the LCD-based designs would only be true if you had a very small point-source lamp. But we don't use $1000 lamps! We use $25 to $60 US lamps with 16 to 25 mm long arcs. So it just never occurs that "Each lens small area working on a specific LCD area".
If you remove the triplet and put a piece of white paper where the triplet goes, then you can see what I mean. Use your computer to generate an image for the LCD with a very small circle of white in the center surrounded by all black. You will not see a small circle of light on the paper! Because of the large lamp arc, you will see a much larger circle on the paper. Every part of the lamp arc will shine through a single pixel and spread that light over the entire size of the projected arc image.
If the lamp arc is 24 mm long and you use 220 and 550 mm fl fresnels, then you will get around a 60 mm arc image through every LCD pixel:
arc image size = 24 * (550 / 220)
If you use the right focal length fresnels, you can make all of those 60 mm circles fall at the center of the projection lens. If you use longer fresnels, then those 60 mm circles will each use an area from the center of the lens all the way to the outer diameter of the lens. Since the center area of a lens gives you a much higher image quality than the edges, longer focal length fresnels will give you an image that is not as sharp. If you use even longer fresnels, then some part of the 60 mm circles from the LCD corners will fall past the first lens surface, so the corners will be dim.
Yes Guy Grotke, my example is ideal (0 arc lengh) but even if you consider 24mm arc length and 60mm projected arc lengh inside the triplet, then "each small lens area will work on a LARGER lcd SEMI-exclusive area but will be still true overall, mean that the botton side of the lens will never work for upside of the lcd. Lets say it will be midleway between difuse light and perfect fresnell based setup, do you agree?
And the 1cm outer circle sement that i work out is not working in your setup, is based on the optical center focusing teory and 60mm arc lengh inside it. IF you trace the rays to 17" corner, you will not use full lens aperture. If i did this with 0 arc lengh, then the unused area would be much larger. Need a pic?
600 mm focal would project 65mm arc lengh and there wouldn´t be unused area in the rear lens. (neither would be light out of the triplet teorically. I´ll try to do some pic later...
And the 1cm outer circle sement that i work out is not working in your setup, is based on the optical center focusing teory and 60mm arc lengh inside it. IF you trace the rays to 17" corner, you will not use full lens aperture. If i did this with 0 arc lengh, then the unused area would be much larger. Need a pic?
600 mm focal would project 65mm arc lengh and there wouldn´t be unused area in the rear lens. (neither would be light out of the triplet teorically. I´ll try to do some pic later...
An externally hosted image should be here but it was not working when we last tested it.
lease check my maths. This is upside half setup with 17", 20mm gap, 550 fresnll and 135 triplet, for 100" image. (triplet center optical not at center check it).
The arc lengh is 60mm so 30mm would be upside at 550 from the fresnell. if you draw the lines, you´ll see there is very small circle in the outer of the triplet that is never used (you can paint it flat black and nothing would change on the projected image). It is 12mm circle segment that is not being used because of the field fresnell focal of 550mm.
here is a pic of rear lens use;
An externally hosted image should be here but it was not working when we last tested it.
i know it is very small area and that hardly would be apreciated, but if we are designing the ideal fresnell, i found it is not ideal 550. Any coments?
Rox you have disproved your own theory. In your pics in post #70 you have drawn the two extremes of the way a lens can focus. One with a diffuse source and the other with a perfect point source. The real world system is in-between with a medium light source (16-25mm arc). In the diffuse source the entire lens aperture is being use to collect light and focus it. In the perfect point source only small points of the lens are being used to collect light and focus it. With the 16-25mm arc, only parts of the lens are being used for each pixel. The entire lens is not being used and small points are not used either.
In your post you say by blocking half of the lens of the diffuse source that the image improves. You then say ” (Why better? think that the other half will be hiperprjected on top, so litlle error will be between them that will become on less sharpness).” What you have explained is that by blocking half of the lens you have effectively made the lens smaller and you have less distortion and aberration effecting the final image. This is the exact reason why not to use the full lens aperture. By lighting the entire lens aperture with the arc, as you are suggesting, you have effectively made the lens collect light the same way as the diffuse source example. The correct way is to use less of the lens aperture, focusing light to the centre of the lens, this ensures that all the light gets through and aberration and distortion is kept to a minimum.
Here is another example. Photographers regularly use larger lenses that are stopped down, reduced in size, so that they can capture a better quality image. The down side of this is that the photographer has to increase the exposure time or use different film.
DJ
In your post you say by blocking half of the lens of the diffuse source that the image improves. You then say ” (Why better? think that the other half will be hiperprjected on top, so litlle error will be between them that will become on less sharpness).” What you have explained is that by blocking half of the lens you have effectively made the lens smaller and you have less distortion and aberration effecting the final image. This is the exact reason why not to use the full lens aperture. By lighting the entire lens aperture with the arc, as you are suggesting, you have effectively made the lens collect light the same way as the diffuse source example. The correct way is to use less of the lens aperture, focusing light to the centre of the lens, this ensures that all the light gets through and aberration and distortion is kept to a minimum.
Here is another example. Photographers regularly use larger lenses that are stopped down, reduced in size, so that they can capture a better quality image. The down side of this is that the photographer has to increase the exposure time or use different film.
DJ
Dazzzla said:
Yes i agree that it is midleway between perfect 0 arc lengh and difusse source. The top part of lens on the fresnel based setup, will never work on the botton lcd image. There is some lens area that will work on some area on the lcd, but FULL lens area will work on lcd if we add all "small works".
If i block the light from the difuse source (block upside), do you agree that the oposite side will be overall better quality? i mean don´t you think the lens upside was doing better work on upside than downside of the image before i stoped the light? I mean the distortion introduced by the lens is not simetrical. So so when we have fresnell based setups, overal the distorsion is minimal, Why do you think there are singlet OHP? The resulting image has nothing to do with difuse sources and a singlet.
I believe this thing i am calling "small lens area working on small lcd area" is much important than the aberrations introduced by the lens when you are on the outer limit.
If it was like you say, that we should keep close to the center so we keep the distortions minimal, then why are you focusing so far on the optical center? you can focus exactly to the center of the rear lens so minimal area would be used and "minimal" aberrations would be introduced. A 60mm diameter rear lens would do a perfect work to you.
You originally wanted to know why we need to focus the light to the centre of the lens so that you have a basic theoretical modal. With a dcx lens for example there is two refracting surfaces. So by focusing light to the centre of the first surface, the second surface still influences the quality. If a pcx lens were used as an example, the optical centre for this would be the plano side of the lens. A triplet is just multiple lenses so the centre of the combined elements is used as a rough guide as to its optical centre.
DJ
Dazzzla said:
yes, i still would like to know. 😀
i believe optical center is nor real. There is noting there. Is just a simplification so maths are more easy and manageable.
I still hold that focusing at full rear lens is ideal (but it is true that the arc lengh needs to be considered and that the outer ray needs to be inside the rear lens as well, i agree 100% with it)
I would like to listen, why focusing the optical center is the ideal setup, no care the lens size, shape, diameter... only focus the optical center.... don´t understand. I can give you an example if you want where focusing optical center will not work.
yes
no
It's also the reason why a triplet is better.
Rox I think you are missing our point. By using a lens in photography, Why is the image better when the lens is stopped down(made smaller)?
DJ
Although most of your ideas will work, you want to find a basic modal. When I say that the optical centre out the lens should be used to focus the light, this is so that the modal fits all set-ups. Yes filling the entire lens aperture will work with the 135mm FL triplet and yes filling the entire lens aperture of a singlet will work. But filling the entire lens aperture of all lens types wont work and more aberrations will be present
DJ
DJ
- Status
- Not open for further replies.