Why a pentaprism?

[denizenx]

OT

Air Lens
The air spaces between the glass lens elements making up a photographic lens can be thought of as lenses made of glass having the same index of refraction as air (1.0). An air space designed from the beginning with this concept in mind is called an air lens. Since the refraction of an air lens is opposite that of a glass lens, a convex shape acts as a concave lens and a concave shape acts as a convex lens. This principle was first propounded in 1898 by a man named Emil von Hoegh working for the German company Goerz.

but damn cool nonetheless

 

[roygoh]

Originally posted by ckiang
AFAIK, lenses ALWAYS produce upside down, laterally inverted images. Rule of optics.

In the most common SLR diagram above.

1. Mirror upside down, l/r inverted on mirror.
2. Image rightside up but l/r inverted on the TOP of the screen.
3. After 1st reflection on the pentaprism, rightside up, l/r inverted.
4. After 2nd reflection, the inversion is corrected and passed to eyepiece.


Regards
CK

Aiyah, ckiang, you made me stay another 30 minutes more in office creating the picture below, which is a 3D visualisation of the pentaprism shown in the picture that denizenx has shown. I have omitted the lens as well as the reflex mirror that is normally at 45 degrees between the film plane and he focusing screen.

I am pretty sure that my representation would be correct if SLR pentaprisms are really as depicted in denizenx's diagram. In that case the image will be l/r inverted. Let me know if you spot any mistakes in my interpretation.

So, the mystery remains unsolved, for me at least......

- Roy

 

[ckiang]

Got me all confused.

Maybe prisms work differently from mirrors?

Regards
CK

 

[denizenx]

Originally posted by ckiang
Got me all confused.

Maybe prisms work differently from mirrors?

Regards
CK

same lah, for pentaprism... the flat surfaces are all mirrors, w or wo light loss...

 

[ckiang]

Where are the experts when we need them? Ian? Jed?

Regards
CK

 

[denizenx]

Originally posted by ckiang
Where are the experts when we need them? Ian? Jed?

Regards
CK

I thought Ian has voiced his expert knowledge already?

 

[roygoh]

Originally posted by ckiang
Got me all confused.

Maybe prisms work differently from mirrors?

Regards
CK

You mean my diagram is confusing? How can you say that, after all the effort I put in to creating that diagram! Tell me first whether or not you agree with my diagram.

Anyway, each reflecting surface works just like a mirror. Unless there are curved surfaces, 2 reflections as shown will not present a L/R correct image to the user.

Originally posted by ckiang

Where are the experts when we need them? Ian? Jed?

Yeah, where are they?

 

[ckiang]

Originally posted by roygoh


You mean my diagram is confusing? How can you say that, after all the effort I put in to creating that diagram! Tell me first whether or not you agree with my diagram.

Anyway, each reflecting surface works just like a mirror. Unless there are curved surfaces, 2 reflections as shown will not present a L/R correct image to the user.

Yeah, where are they?

No, I mean I have never analyzed it to this level. Yours and denizen's look correct, all the SLR diagrams are that way. But doesn't appear to be like that in real life. Something is amiss...

Regards
CK

 

[roygoh]

Here's what I managed to find. The representation of the pentaprism in the diagram beloiw is clearly different from the simplistic representation of most other diagrams.

Hope that the link works. This is where I got the picture from:

http://www.profotos.com/education/promag/articles/september2001/modernslr/index.shtml

 

[denizenx]

Originally posted by roygoh
Here's what I managed to find. The representation of the pentaprism in the diagram beloiw is clearly different from the simplistic representation of most other diagrams.

Hope that the link works. This is where I got the picture from:

http://www.profotos.com/education/promag/articles/september2001/modernslr/index.shtml

u mean the diopter eye piece could be an inverter as well? NOOooooooo........!
but could be leh, the cheapest and stealthiest way to do it... and leave all the reverse engineers eating their hats...

 

[roygoh]

Originally posted by denizenx


u mean the diopter eye piece could be an inverter as well? NOOooooooo........!
but could be leh, the cheapest and stealthiest way to do it... and leave all the reverse engineers eating their hats...

That's not what I meant. If the diopter optics is design to invert then the image will be up-side down again...

What I was referring to is the shape of the pentaprism shown. It is more complex than the normal representation in your diagram, which I based my digram off.

I believe the largest surface on the top shown is the first reflection surface, the second reflection surface is on the other side which cannot be seen, and the the surface on the right which is directly in front of the diopter optics is the 3rd refletion surface.

What do you think?

 

[denizenx]

Originally posted by roygoh


That's not what I meant. If the diopter optics is design to invert then the image will be up-side down again...

What I was referring to is the shape of the pentaprism shown. It is more complex than the normal representation in your diagram, which I based my digram off.

I believe the largest surface on the top shown is the first reflection surface, the second reflection surface is on the other side which cannot be seen, and the the surface on the right which is directly in front of the diopter optics is the 3rd refletion surface.

What do you think?

could be could be.... that means besides flipping up and down there's a intermediate left right flip... but very jialat to engineer cos there's mucho distortions..
but then it's septaprism++ liao... ok me nitpicking
as usual my solutions super simplistic hehe cos use cylindrical diopter will just LR the RL tho image quality sux..


SIGH can someone just take out his F5?????

 

[roygoh]

Originally posted by denizenx


SIGH can someone just take out his F5?????

....and take a picture of the "Pentaprism" to show us all here.:gbounce:

 

[ckhaos]

Hello roy,

I looked at your diagram long and hard, and I think I know what is the problem now. Total internal reflection by a prism surface causes inversion as well, which you did not factor into your diagram. Here is a link to an amended one:

http://www.pbase.com/image/11717130

The object altogether undergoes 4 inversions -- first through the lens, second through the reflex mirror, and third and fourth through the prism surfaces. Thus, in the end, the object is upright and lateral.

 

[ckiang]

Originally posted by ckhaos
Hello roy,

I looked at your diagram long and hard, and I think I know what is the problem now. Total internal reflection by a prism surface causes inversion as well, which you did not factor into your diagram. Here is a link to an amended one:

http://www.pbase.com/image/11717130

The object altogether undergoes 4 inversions -- first through the lens, second through the reflex mirror, and third and fourth through the prism surfaces. Thus, in the end, the object is upright and lateral.

Still not right. Image on the focussing screen should be right side up and laterally inverted. The image there is transmitted via a mirror.

Regards
CK

 

[roygoh]

Originally posted by ckhaos
Hello roy,

I looked at your diagram long and hard, and I think I know what is the problem now. Total internal reflection by a prism surface causes inversion as well, which you did not factor into your diagram. Here is a link to an amended one:

http://www.pbase.com/image/11717130

The object altogether undergoes 4 inversions -- first through the lens, second through the reflex mirror, and third and fourth through the prism surfaces. Thus, in the end, the object is upright and lateral.

Alamak, your diagram totally wrong lah. When you look into a mirror your reflection is still upright though LR inverted right?

How can a TIR cause both a LR and and up-down inversion at the same time?

The first inversion through the lens is both up-down and LR. The reflection from the reflex mirror creates an image on the focusing screen that is upright if you view from the top of the camra but LR inverted. The 2 reflections shown can only bring the image upright towards the rear of the camera but not correct the LR inversion.

Try laying a picture on the table, upright when you view at it from the top. Next take 2 mirrors and try to simulate the 2 reflections such that you can see the image of the object looking from a direction parallel with the table top and you will see an upright and LR correct image.

The problem in the SLR case is that the image on the focusing screen is already LR inverted.

 

[Peculiar]

Originally posted by roygoh


Aiyah, ckiang, you made me stay another 30 minutes more in office creating the picture below, which is a 3D visualisation of the pentaprism shown in the picture that denizenx has shown. I have omitted the lens as well as the reflex mirror that is normally at 45 degrees between the film plane and he focusing screen.

I am pretty sure that my representation would be correct if SLR pentaprisms are really as depicted in denizenx's diagram. In that case the image will be l/r inverted. Let me know if you spot any mistakes in my interpretation.

So, the mystery remains unsolved, for me at least......

- Roy

Actually your image is partially wrong.. the top of the prism is actually pointed, such that the first reflecting surface is like a roof shape, giving is 2 reflecting surfaces.. This is where the L/R inversion correction takes place, before it hits the next reflecting surface to make the image upright..
Quite hard to visualise, so sorry for lack of diagram..

Conventional prism diagrams show only the cross-section at the highest point of the prism, so doesn't indicate the "roof top"

I noticed the actual shape of the prism when i took apart of my Nikon F3 prism head... ;p

 

[roygoh]

Originally posted by Peculiar


Actually your image is partially wrong.. the top of the prism is actually pointed, such that the first reflecting surface is like a roof shape, giving is 2 reflecting surfaces.. This is where the L/R inversion correction takes place, before it hits the next reflecting surface to make the image upright..
Quite hard to visualise, so sorry for lack of diagram..

Conventional prism diagrams show only the cross-section at the highest point of the prism, so doesn't indicate the "roof top"

I noticed the actual shape of the prism when i took apart of my Nikon F3 prism head... ;p

I get what you mean. I know there has to be something wrong with the simplified representation of the pentaprism shown in most diagrams. My image is wrong because it gives an LR inverted image which is clearly not the case in SLRs.

Is the roof-top shape equivalent to 2 mirrors at 90 degrees to each other? When you look into a setup of 2 mirrors at 90 degrees, your image is LR correct because it has gone through 2 reflections. However, the joint between the 2 mirrors will also be visible. That could be caused by poor joining, which will not be the case if the reflecting surfaces are from a prism.

I am still slightly doubful why the intersection line of the roof-top is not visible.

Thanks for your information. Any chance you can show a picture of your F3 prism?;p

- Roy

 

[Peculiar]

Originally posted by roygoh


Is the roof-top shape equivalent to 2 mirrors at 90 degrees to each other? When you look into a setup of 2 mirrors at 90 degrees, your image is LR correct because it has gone through 2 reflections. However, the joint between the 2 mirrors will also be visible. That could be caused by poor joining, which will not be the case if the reflecting surfaces are from a prism.

I am still slightly doubful why the intersection line of the roof-top is not visible.

Thanks for your information. Any chance you can show a picture of your F3 prism?;p

- Roy

Whether the 2 mirrors at 90 deg, i'm not quite sure, but the prism in my F3 is around 90 deg... This i think may differ between different cameras.

As for the invisible intersection line, i think the use of oversized prisms helps in concealing them...

A picture? I'm afraid i can't post it up here, because i don't have access to any scanner or digital cams... Perhaps some other time, or anyone out there willing to take apart their F-series Nikon camera prism heads?? ;p

 

[chriszzz]

Originally posted by roygoh


I get what you mean. I know there has to be something wrong with the simplified representation of the pentaprism shown in most diagrams. My image is wrong because it gives an LR inverted image which is clearly not the case in SLRs.

Is the roof-top shape equivalent to 2 mirrors at 90 degrees to each other? When you look into a setup of 2 mirrors at 90 degrees, your image is LR correct because it has gone through 2 reflections. However, the joint between the 2 mirrors will also be visible. That could be caused by poor joining, which will not be the case if the reflecting surfaces are from a prism.

I am still slightly doubful why the intersection line of the roof-top is not visible.

Thanks for your information. Any chance you can show a picture of your F3 prism?;p

- Roy

OK, I have an old Canon F-1 which has a removable pentaprism, and I removed it's pentaprism, and I looked UP into it, just as the light from the SLR reflex mirror would. I can see my own reflection upon tilting it, and it is laterally inverted ( L<->R ).

Going in the opposite direction and looking through the viewfinder of the pentaprism also reveals a laterally inverted ( L<->R ) image. So this lateral inversion counters the lateral inversion from the lens, and so, turns the image the right way.

Looking up into the pentaprism reveals 2 slanted surfaces ( ie, the sides of the prism ) that act as the reflecting surface. As Peculiar as pointed out, these 2 surfaces join together without a visible line. Remember that this is a high precision optical component, so it's no surprise that line is not visible (perhaps it might be visible under high magnification, but I could not see it under normal indoor lighting).

The left surface reflects the right half of the image, while the right reflects the left. The corrected image is then reflected to the front of the pentaprism, which reflects it back to the viewfinder into your eye.

There is a picture of the Canon F-1 here : http://www.ne.jp/asahi/japan/manual-camera/f-11.htm . The shape of the pentaprism is very obvious. The sides are slanted, which are the 2 reflecting surfaces mentioned.

Hope this clears it up at last.