View Full Version : what to use as a relay lens? + my 35 adapter concept


Rok Furman
January 15th, 2006, 03:00 PM
i'm toying with some concepts and thinking of constructing my very own 35mm adapter to fit on a XL-1.

this is a concept sketch of the prototype:
http://users.volja.net/furmanr/das_35.gif

i'd use the same concept as in a single lens reflex camera - the light from the lens would be projected onto the ground glass, which is oscillating or spinning, via a mirror.
the projected image would be then rotated a 180° within the roof pentaprism and a "macro" lens would be used behind the pentaprism (not drawn, picture it where the arrow point is). the light would continue it's path onto the XL's sensors - PRESTO! - the entire 35mm frame on a 7.2 times smaller area.

i know it works, because i looked through the viewfinder of an EOS i use, with the XL lens attached and the picture was oriented as it shoud be.

it would probably need a condenser and acromat as well, but i have one major unanswered question; what to use as the relay (macro) lens?
i'll try with a 50mm M42 thread lens which will require an adapter which i intend to mill myself;

http://users.volja.net/furmanr/mount-1.gif
http://users.volja.net/furmanr/mount-2.gif

an original XL bayonet fitted to a barrel with an M42 thread at the end. if i were to use only this combination i would get a 360mm lens (50mm * 7.2 crop). cheap so far.
but this lens doesn't offer the quality you get out of the kit lens:

http://users.volja.net/furmanr/DDR1.jpg
http://users.volja.net/furmanr/L.jpg

the later being taken with the standard XL lens, the first with the 50mm M42 lens, which i held to the bayonet with my bare hands :)
i would certainly need macro bellows to get the distance right. i might have to use a prime lens with a shorter focal lenght, as the 50m could make my camera with the adapter installed look like a rocket launcher.

another problem is with the image quality. the 50mm isn't as good as the kit lens, and the pentaprism would also distort the image - more glass, more problems. i could use the XL lens as the relay lens, but it has A LOT of glass in it and i want to avoid that. also, i would have to use + macro adapters to get the lens close enough to fit the entire 35mm frame in it.
i could also do it without the pentaprism and mirror and project the image directly on the ground glass, and use the prism in an extention of the viewfinder, so that i could composite the shot - and then rotate the recorded image in post.

so, the main questions from me would be:

is my concept any good?

could i use the 50mm lens (or some other prime) as the relay lens, or should i rather use the XL kit lens with + converters? something else?

should i use the pentaprism to view the image in the viewfinder correctly oriented or should i bypass it and have an upside down image in my viewfinder? the prism would probably be from an old medium format camera or hand-crafted

keep in mind that i want to keep the adapter as light and practical as possible. i would mount old M42 or Canon FD prime lenses on it. any fully manual lenses, basicly... the bayonets aren't that hard to mill.

Bob Hart
January 15th, 2006, 10:52 PM
The pentaprism might be the weak link. It might have to be larger than a viewfinder prism to avoid picking up edges or getting falloff on edges. I think well back in the AGUS35 or ALDU35 threads somebody found success with camera pentaprisms on very small videocameras.

For larger 3CCD cams with 58mm or 72mm filter mounts, I don't think you'll get the entire image in.

I can't see why the overall concept can't work providing that is a surface-coated mirror you intend to use.

Are you using a three-row or four-row peripheral ballbearing eccentric to move the groundglass. If so, you may find you have a noise or a chatter problem if you are using the pillars and 1mm clearances to prevent the orbiting gg from spinning and leaving you with a grainy centre in the image.

One solution is to use another drive belt or better, two for balance of forces to limit the effects of wear, allowing the gg frame to float and use the pair of belts onto a fixed (not moving) pulley shape to stabilise the gg. The pre-load from the belts should be constant and the appliance quieter as a result.

As for a relay lens, I suggest you start with a 25mm or thereabouts C-mount security camera lens, as small in the barrel diameter as you can find it. Position it about 11mm furthur foward of the focal plane than it should be, ie., 28.5mm v/s 17.5mm. This should yield sharp focus on a 24mm x 18mm frame at about 120mm distance from the lens. This arrangement may put the flange face almost on the same flange plane as the XL1 lens.

Quyen gave very good advice on experimenting with XL1s. Makes sure any adaptor you make does not come into contact with the electrical conductors in the lens mount otherwise your camera will be ruined.

Don't do anything for a while until others post here to challenge my comments as I am not an engineer of any sort.

Quyen Le
January 16th, 2006, 01:24 AM
Bob gave very good advice there. There are many things to consider when trying some lenses for relay lens.
- IRIS. Without this, you will suffer shallow DOF when shooting in good lighting. Many premade relay lens at Edmund optics doesn't have this. The first P+S version is believe not to have iris control on the adapter. However, ND filter can be used instead to get good shallow DOF.
- Image size. You want 24x36, 18x24 or smaller? the larger the image, the bulkier the adapter.
There are some lenses out there, but to get the best out of your XL1, buy P+S relay lens if you have the fund.
My Letus35XL is on the way and won't be out for at least another 2 months. I will get some footage when I have time and will upload for you all to judge. I am pleased with the results so far. Thanks.

Quyen

Rok Furman
January 16th, 2006, 09:41 AM
Bob, what did you mean with "surface-coated mirror"? what's the difference between that one and a "normal" one? :)
i'll probably rotate the GG instead of vibrating it. or something. i need a full concept first :D

how much does the original P+S relay lens cost, Quyen? i'm working on this adapter as a hobby, i don't have much money to invest.
another question: what would be the flange distance for a XL mount i.e. how far are the CCDs from the rim? i hope you can provide the answer.

the nominal flange distance for a M42 lens is 45,5mm, which is labled as "B" on this sketch (the "senzor" being the CCDs or film):
http://users.volja.net/furmanr/mount-4.gif
what i want to know is how far is the rim of the XL mount from the CCDs (the distance labled "A")

in other words... i want to know the lenght the "barrel" of the adapter - compare this sketch to the previous ones if you have trouble

Bob Hart
January 16th, 2006, 11:46 AM
A normal mirror is a piece of glass with a reflective coating on the back and a thick layer of redlead to stop it from getting scratched off or going black.

A surface coated mirror has the reflective coating on the front.

A normal mirror is fine for examining zits and deciding whether that hair transplant is needed just yet. Try to use it in fine optical applications and you will get multiple images (like the most maladjusted VHS you ever saw) a sort of ghosting effect if you can remember back to analogue TV.

This is caused by the image reflecting off the back layer, bouncing back onto the back layer from the front surface then maybe back and forth a few more times depending on the brightness of the subject.

If you use normal mirrors to erect an image you end up with this multiple pictureframe effect - not very nice.

Surface coated mirrors are the only way to go. This is what is used in SLR cameras. They are more prone to damage and have to be treated very carefully.

I think I have seen reference to the Flange to focal plane on the XL1 in older threads on the XL1, possibly related to the static devices.

Bear in mind that 3CCD cameras use a prism block to split the image into three colour zones (my paraphrase). The thickness of this prism determines that the focal plane (CCD) would be physically furthur back than the focal plane the videocamera lens uses.

My understanding is that to compensate for this, there are other optics between the prism and CCDs. I think this may be the on-chip lenses referred to in some brochures.

The "A" distance may well be longer than the distance "seen" by the XL1 lens because the on-chip lenses may bring the CCDs optically closer to the XL1 lens than it actually is in real distance.

The "in-air" distance should be easy enough to find out. Just take the lens off your XL1 and set it up in a vee block or a jig, then get a piece of groundglass, lunchwrap or a plastic cap from a small pringles tin.

My original AGUS35 was a Pringles can, the camera lens jammed up one end wrapped in a rolled up sock. The end of the sock and the metal bottom of the can had both been removed of course.

Some of the small cans have a lid which is clear plastic and has the frosted finish added on the inside afterwards. I think these originate in Europe whereas the big ones come from the US.

Using your substitute groundglass as an image plane and looking by eye at the image on this groundglass, move it back and forth, closer to or furthur from the rear of the Canon lens until you get the sharpest image.

This will give you an approximate flange to focal plane when you measure the distance. For most accurate result, focus the lens on infinity and point it at a very distant object. Power pylons on top of hills are quite good.

Looking at your diagram, it may not work as the M42 to bayonet fitting adaptor may put the relay lens too far forward. - But I am only guessing here. My imagining of the relay lens is that part of it may have to sit back inside the space inside the XL1 mount if that lens is a C-mount lens. The M42 lens may not be able to get back in far enough because it and the adaptor it screws into, may not fit within the camera body inside the XL1 lens mount.

If you want to get a rough idea of what you would get with a smaller relay lens, set your XL1 lens on about the 25mm position on its zoom range with the focus on infinity, disconnect it from the camera and bring it forward about 11mm. Look in the viewfinder or patch out to a larger TV screen.

A patch via firewire into a computer into your NLE on capture so you see the full frame, would be better as this is effectively an underscanned image.

You'll have to have a vee block or jig and have the centreline of the lens lined up with the centreline of the camera for this to work. You may also have to make a crude tube of cardboard or thick paper for between the back of the lens and the camera to keep the stray light out.

Get an envelope with a postage stamp on it or a grocery container with a barcode on it - move this to about 5 -7 inches away from the front of the lens and see if this produces a sharp image in the camera at full framing of the stamp or barcode.

The XL1 might not work with the lens off.

This trick may only give you a broad hint of what might work. Zoom lenses behave differently to prime lenses when set forward like this. The results from different models of zoom lenses are also quite varied.

If you bench test like this, make sure you prevent the camera lens from rolling off onto the floor.

This is my vague understanding and as I mentioned previously, wait awhile and let others wield the stick and break apart the possibly corrupt pinada my own info may be so that it might be appropriately consigned to rthe waste bin of bad ideas.

I know an independent Australian film-maker whose ancestry traces back to your country. His name is Frank Rijavec, notable for a feature documentary "Exile and the Kingdom".

Rok Furman
January 16th, 2006, 03:13 PM
this has got to be the best post i've ever read on any forum! i thought it had to be something like that with the mirror, i was also considering to put a polished stainless steel plate or something
you've clear up a lot in my head, thanks. i'll wait for others to reply.

the XL kit lens is an automatic lens, which has to be connected to the camera in order to work. the camera powers up it's servos, meaning i can't move anything while the lens is off the camera - not even the zoom ring, since power is needed to run it's sensors, which are required to move the optical group... even the iris is closed shut while the lens has no power.

the camera, however, can work perfectly fine without the lens - i might try to catch the flange distance for the M42 lens by feel as i did in the frame grab (DDR1.jpg) in my first post.

as for the optics, i'm gonna have to look stuff up, i didn't attend physics classes when we learned about lenses and mirrors very often, almost cost me my year too :P

thanks yet again for the info and give my kind regards to Frank. tell him to drop me a line if he ever comes to slovenia as a tourist :)

Greg Bates
January 16th, 2006, 05:56 PM
I found this link before I discovered that I was optically challenged. Also somewhere in here Matteo (who is well known for hacking the lens off his z1u) mentioned that he used a 24mm reflex macro lens for his relay. I'll find the thread and put the link in.

http://www.naturfotograf.com/roll_your_own_lens.html

Michael Maier
January 18th, 2006, 01:52 AM
Also somewhere in here Matteo (who is well known for hacking the lens off his z1u) mentioned that he used a 24mm reflex macro lens for his relay.

I don't think a SLR lens will get you good results.

Greg Bates
January 18th, 2006, 08:07 AM
I don't think a SLR lens will get you good results.
How soon we forget Michael! This was your favorite footage!

http://wave35.altervista.org/

Michael Maier
January 18th, 2006, 09:44 AM
You are right Greg, but my enthusiasm was mainly based on the footage from the girl only, which was very brightly lit and almost washed out or bleached if you will. The high contrast and brightness can hide a lot. After taking a closer look at the darker footage, including looking at it in a HD monitor and playing with it in CC and different resolutions, I saw a lot of aberrations not present for example in the G35, SG35 or Letus35 footage. I'm thinking it may be the SLR lens in the set up. Also remember all those clips seem to be heavily color corrected.

Greg Bates
January 18th, 2006, 10:17 AM
I can't argue there, for as i've said I am not optically inclined enuff to try it. I wish I was. But what really is a relay lens then? I mean I figure with the xl2 you would need something that had a very short focal length when you consider the 7x factor on the ccd. so perhaps the other link would be viable for an xl? I don't know it gives me a headache...I leave it to the gurus.

Rok Furman
January 22nd, 2006, 03:53 AM
i was thinking... how exactly does a director's viewfinder work? you know, the things you see spielberg lugging around the set, that you can mount a PL lens in the front... i could incorporate this concept in my project.

it probably has a ground glass in it, but how does it flip the image?

a photo of the device:
http://www.cavision.com/viewfinders/VFC52.jpg

and a shorter one:
http://www.cavision.com/viewfinders/VFPL52.jpg

Michael Maier
January 22nd, 2006, 03:59 AM
Most likely the same way a SLR camera's viewfinder does.

Rok Furman
January 22nd, 2006, 04:06 AM
i think not - check out it's construction. it's a lens mount followed by a barrel (or tube)... there couldn't be a pentaprism or mirror in there

Rok Furman
January 22nd, 2006, 04:09 AM
hmm, come to think about it... a Abbe-Koenig or Schmidt-Pechan prism could be inside it.
they are used in binoculars

http://en.wikipedia.org/wiki/Abbe-Koenig_prism
http://en.wikipedia.org/wiki/Schmidt-Pechan_prism

if anybody has any knowledge of the director's viewfinder i'd love to hear it :)

Rok Furman
January 27th, 2006, 06:52 AM
hmm, in general what would produce a better picture: a prism mounted in frount of the relay lens (to turn the picture 180°) or a set of surface coated mirrors that would do the same?

i'm begining to abandon the prism idea and thinking about putting the prism in the viewfinder so i can still composite the shot.

or taking the camera apart and physicly turning the CCDs 180° :D
or taking the viewfinder apart and doing the same :D

Bob Hart
January 27th, 2006, 11:00 AM
Surface coated mirrors should lose less light, but may be difficult for some builders to set up accurately and to keep correctly aligned. They are light in weight and tend less to injury from shock loadings if successfully installed without any structural loads placed on them.

Prisms already have their reflective surfaces set up accurately and are easier to mount. If the reflective faces are coated, then light loss can be reduced. They are however heavy and their mass requires much more robust support and insulating structures around them to cope with shock loads.

Don't give up on prisms or mirrors. Life is a lot easier if the image is upright into the camera.

R.P. Cuenco
January 27th, 2006, 05:11 PM
i already tried rotating the ccds 180 (just for kicks, dont ask why), and the cables are just too short to accomodate a half turn since they're customized.

Rok Furman
January 28th, 2006, 04:41 AM
i'm thinking of using this lens as a relay lens:
http://www.cbcamerica.com/cctvprod/computar/mono/125.html
it has an aperture control ring which will come in handy

i'll probably have a 36*24mm frame, such as in a SLR camera - is this lens enough or should i get a lens with a greater focal lenght?

again, what is the flange distance for the XL mount?


btw is there any other way to turn the picture around without prisms or mirrors?
R.P. Cuenco, i could put in longer cables :D

Joel Kreisel
January 28th, 2006, 06:38 AM
btw is there any other way to turn the picture around without prisms or mirrors?

Check out the thread I just started on that very subject. I have a theory, but no conclusions yet.

Also, there are other kinds of prisms aside from pentaprisms that will do the job. In this case, a dove prism (http://www.optosigma.com/miva/merchant.mv?Screen=PROD&Store_Code=OS&Product_Code=pg141&Category_Code=Prisms+%26+Windows) might be more what you're looking for.

Alain Bellon
January 28th, 2006, 11:56 AM
Check out the thread I just started on that very subject. I have a theory, but no conclusions yet.

Also, there are other kinds of prisms aside from pentaprisms that will do the job. In this case, a dove prism (http://www.optosigma.com/miva/merchant.mv?Screen=PROD&Store_Code=OS&Product_Code=pg141&Category_Code=Prisms+%26+Windows) might be more what you're looking for.

The other prismatic option aside from a pentaprism plus a mirror is a set of two right-angle prisms (porro prisms are optimized versions with rounded corners and a slit in the middle to minimize internal reflections). The right angle prism coupling is the same you encounter inside binoculars. The penta prism + mirror is the configuration found in SLR cameras.

Both of these configurations can be achieved by mirrors alone placed at the same positions where the prism facets are.

Rok Furman
February 8th, 2006, 01:09 PM
the progress so far:

http://users.volja.net/furmanr/35.avi
http://users.volja.net/potocnikda/35a.avi

it's only a concept build so the quality just isn't there... yet :D
the relay lens is the kit XL lens, but i'll replace it with a CCTV lens when i aquire one. i still have to rotate the GG, it's static for now :)