[mkaake]

Just a quick question about an unsplit design. I've followed a lot of projects here, and have read both of the wikis, but there's something that I wanted to verify...

With unsplit optics, do we continue to use 2 fresnels so that we can ensure the angle of the light entering the second fresnel? It really seems like you could do away with the first fresnel in the split design, unless you had the requirement that the light entering the second fresnel was parallel...

Eager to learn...

[Sinner7]

mkaake, on Jul 12 2007, 09:59 AM, said:

Just a quick question about an unsplit design. I've followed a lot of projects here, and have read both of the wikis, but there's something that I wanted to verify...

With unsplit optics, do we continue to use 2 fresnels so that we can ensure the angle of the light entering the second fresnel? It really seems like you could do away with the first fresnel in the split design, unless you had the requirement that the light entering the second fresnel was parallel...

Eager to learn...

With a single, small lamp arch as a light source, you need both fresnels all the time to collect, spread and condense the light through the triplet. If you remove either fresnel, you will lose almost all of your light for projection resulting in a dim, unwatchable image.

Using a LED array could eliminate the rear fresnel because you are flooding the lcd panel with 100s of tiny light sources rather that one lamp. Hope that helps.

[tgreenwood]

mkaake, on Jul 12 2007, 09:59 AM, said:

Just a quick question about an unsplit design. I've followed a lot of projects here, and have read both of the wikis, but there's something that I wanted to verify...

With unsplit optics, do we continue to use 2 fresnels so that we can ensure the angle of the light entering the second fresnel? It really seems like you could do away with the first fresnel in the split design, unless you had the requirement that the light entering the second fresnel was parallel...

Eager to learn...

I think I know what you are saying, and I think that using just a single fresnel would work just fine.

Here's a picture to tell what I'm talking about: (Did it with this webpage, Physlet Animation - Geometric Optics link )

 Single_fresnel.JPG (26.04K)
Number of downloads: 12

I am now ducking behind my computer desk because this is a radical notion.
But I think it'll work.


Tgreenwood

[jonjandran]

No, no, no.

The first fresnel takes the point light source and collimates it and spreads it over the entire lcd panel. The second fresnel takes the light and narrows it back down to fit in the triplet.

Unsplit works because the second fresnel is so close to the Lcd that the light is still collimated and spread out over the entire lcd panel as it passes through the Lcd.

[NinHowFritz]

Actually I think that it would work if you were able to figure the distances correctly. The fresnel would need to be larger than the LCD, just as in an unsplit design.

However, the fresnel would probably not work well this way, because the light coming out of it is not collimated, and the fresnel grooves will probably make dark rings in the projection. The light in an unsplit design is not collimated at the LCD, but it is between the fresnels where the grooves are.

[tgreenwood]

jonjandran, on Jul 12 2007, 10:22 PM, said:

No, no, no.

The first fresnel takes the point light source and collimates it and spreads it over the entire lcd panel. The second fresnel takes the light and narrows it back down to fit in the triplet.

Unsplit works because the second fresnel is so close to the Lcd that the light is still collimated and spread out over the entire lcd panel as it passes through the Lcd.

OK, I admit that I don't get it.

The fresnels focal length determines the angle that the light goes through the LCD, right?

I just don't see that much difference between using one fresnel for an "unsplit" or two fresnels.

 Single_fresnel.JPG (26.04K)
Number of downloads: 6

 Single_fresnel_I_dont_get_it.JPG (27.29K)
Number of downloads: 6

Could you explain it to me? The light at the LCD isn't collimated anyway either way.


Tgreenwood

[ki113r]

tgreenwood, on Jul 13 2007, 12:17 PM, said:

OK, I admit that I don't get it.

The fresnels focal length determines the angle that the light goes through the LCD, right?

I just don't see that much difference between using one fresnel for an "unsplit" or two fresnels.

attachment

attachment

Could you explain it to me? The light at the LCD isn't collimated anyway either way.
Tgreenwood

I think your bottom picture is wrong , because there is no gap between the second fresnel and the lcd in the unsplit design , so the light remains colliminated from the first fresnel. In your picture the light is starting to focus towards the triplet as it passes through the lcd while in real life it is colliminated and just after it passes through the lcd starts focusing. The top picture is correct as it sows that using just one fresnel light is not colliminated and so the final image ends up dim.

[tgreenwood]

Here's the picture from the Lumenlab DIY Projector Guide (the same drawing is in the pro wiki)
 lumenlab_diy_guide.JPG (13.89K)
Number of downloads: 6

Tgreenwood

[DAZZZLA]

It would be great if we could use a fresnel like a solid lens but we can’t, well we could but it wouldn’t work very well in the majority of cases. If we could find a solid lens the size of a fresnel then this could be used without another and simply gather the light and then converge it back down to fit though the triplet. A fresnel can’t be used this way, it has been designed to efficiently operate only when the source or image is at its focal length on the appropriate surface it was designed for. The further away from its FL the less efficiently it works. As NHF correctly pointed out, the fresnel rings will affect the image and darken the edges. You can see this effect if the wrong FL front fresnel is used. For example if you were trying to use a 330 FL fresnel with a 450-500mm FL triplet. It does work but the edges become very muddy and yellow.

DJ

[ki113r]

tgreenwood, on Jul 13 2007, 03:38 PM, said:

Here's the picture from the Lumenlab DIY Projector Guide (the same drawing is in the pro wiki)
attachment

Tgreenwood

Hmm but by looking at that pic it seems as if the top and bottom of the lcd is not illuminated.

[jonjandran]

ki113r, on Jul 13 2007, 05:40 AM, said:

Hmm but by looking at that pic it seems as if the top and bottom of the lcd is not illuminated.

That's why the fresnels need to be larger than the Lcd.

[mkaake]

ki113r, on Jul 13 2007, 06:43 AM, said:

I think your bottom picture is wrong , because there is no gap between the second fresnel and the lcd in the unsplit design , so the light remains colliminated from the first fresnel. In your picture the light is starting to focus towards the triplet as it passes through the lcd while in real life it is colliminated and just after it passes through the lcd starts focusing. The top picture is correct as it sows that using just one fresnel light is not colliminated and so the final image ends up dim.

Hmm.

So the assumption with the unsplit design is that we're still putting colliminated light through the LCD, and this light is colliminated by the first fresnel?

I'm curious why this assumption is valid when both fresnels are on the lamp side of the LCD?

Also, if the unsplit design places the collecting fresnel next to the LCD anyway, what's the difference with having rings show up in the final projection?

I'm sure you guys are right, and it's been tried before, I'm just having a hard time wrapping my head around it...


<edit>

Quote

The light in an unsplit design is not collimated at the LCD, but it is between the fresnels where the grooves are.

So basically, if the light isn't colliminated when it enters the collecting fresnel, the angle the light is approaching will cast shadows off the grooves of the fresnel? I guess that would make sense, though I'd be curious how bad the effect is, and how it changes with different distances from the light source...

This post has been edited by mkaake: 14 July 2007 - 03:01 AM

[jonjandran]

mkaake, on Jul 13 2007, 10:57 PM, said:

Hmm.

So the assumption with the unsplit design is that we're still putting colliminated light through the LCD, and this light is colliminated by the first fresnel?

I'm curious why this assumption is valid when both fresnels are on the lamp side of the LCD?

Also, if the unsplit design places the collecting fresnel next to the LCD anyway, what's the difference with having rings show up in the final projection?

I'm sure you guys are right, and it's been tried before, I'm just having a hard time wrapping my head around it...
<edit>
So basically, if the light isn't colliminated when it enters the collecting fresnel, the angle the light is approaching will cast shadows off the grooves of the fresnel? I guess that would make sense, though I'd be curious how bad the effect is, and how it changes with different distances from the light source...

Yes the light is not 100% collimated, but once again the fresnels are close enough to the Lcd that the light is collimated enough to work effectively.

If you think about any light like say even an Led. If you put it behind the Lcd then the light spreads out at angles as it moves forward so even that light wouldn't be collimated.

Any light will be not perfectly collimated as it goes through the lcd, but it will be good enough for the most part.

[spankdamonkey]

Sorry to jump in here but i am interested as well, I am in the process of an unsplit design too. My question is why is there so many yes' and no's about an unsplit design? If you do the Overhead projector with a lcd panel it works fine. How come that works?

Thanks

This post has been edited by spankdamonkey: 14 July 2007 - 03:17 AM

[jonjandran]

spankdamonkey, on Jul 13 2007, 11:16 PM, said:

Sorry to jump in here but i am interested as well, I am in the process of an unsplit design too. My question is why is there so many yes' and no's about an unsplit design? If you do the Overhead projector with a lcd panel it works fine. How come that works?

Thanks

It's not a bunch of yes and nos. Both will work fine, but both have pros and cons in some designs.

Even in the professional Ask Impression projectors like the 880 they used an unsplit design at first and then went to a split design for some reason.

[spankdamonkey]

I guess it all comes down to trial and error and what ever works best for the user

[jonjandran]

spankdamonkey, on Jul 13 2007, 11:34 PM, said:

I guess it all comes down to trial and error and what ever works best for the user

Exactly. I am a firm believer in unsplit. It is brighter and better looking in my opinion. But in my Ask 960 retrofit, I had to use a split design because of the uneven blacks. So of course that was best in that particular situation.

There is no 1 perfect solution.

[spankdamonkey]

jonjandran, on Jul 13 2007, 11:48 PM, said:

Exactly. I am a firm believer in unsplit. It is brighter and better looking in my opinion. But in my Ask 960 retrofit, I had to use a split design because of the uneven blacks. So of course that was best in that particular situation.

There is no 1 perfect solution.

Cool i am going to attempt and unsplit with CFL's I am waiting on my portable DVD player to strip and try. I have seem some decent results. I hope to improve on it and see what we can make out of this