macro photography: generating a plane of light?

[dh003i]

At the advice Dan Fromm, I recently bought Nikon Macro-Nikkor 55/2.8 and Tokina AT-X 90/2.5 lenses for use in a macro projects (they're currently at SK Grimes being adapted to a 4x5 lens-board and MP4 polaroid shutter).

I'd like to use these lenses for use in scanning light macrophotography.
Attachment 44552

Here are some diagrams of the SLP setup, which requires creating a plane of light on the sample, and rising the sample through the plane of light as the film is exposed.

Attachment 44553

Attachment 44554

I have a few projectors, but I am thinking that it might be easier to setup the plane of light with a circular ring of light around the sample (and it will also create less harsh light). My two ideas are:

(a) Use a circlight fluorescent bulb and some kind of mylar foil around it to shape the light and block off all but a slit of light along a plane.

(b) Using a DIY fiber optic illuminator made from a CFL, reflector, and a bunch of fiber-optic cables aligned at the light-source, with the fiber optics arranged around a circular disc at the terminal end. See the illustration below:

Attachment 44555

Would either of these work for creating a plane of light? Or would the light diffuse too much by the time it reaches the sample? The fiber-optic cables would be sandwiched between two black flat circular donut-shaped pieces of material. Maybe they could be recessed in inwards a little, so that the light waves would have to pass through a narrow black cave slit after exiting the fiber-optics, hence eliminating off-angle light and making the light directional? Thoughts?

[Mark Stahlke]

Whoosh!
Something just went way over my head.

[dh003i]

Mark, read the scanning light macrophotography article I linked to to get an idea as to what is going on. A summary of the idea is thus:

1. The optimum plane of focus is just that -- a plane, infinitely thin.

2. Depending on CoC used, there is a region of acceptable focus above and below the plane of optimum focus.

3. For 3x macrophotography using a symmetrical lens, assuming a CoF of 0.03mm, the acceptable dof will be say 0.11mm at f/4 (TDF = 2Cf * (m+1)/m^2). DOF can be increased slightly by increasing f#, but you have to deal with more diffraction.

4. Hence, if you want to get an entire sample -- say a small anesthetized frog -- in perfect focus, you could put it on a motorized stage, shine a plane of light parallel to the film plane* on the frog, and raise the sample through the plane of light (via the motorized stage) as he exposure is taken. You would leave the camera focused on the plane of light, so all that it sees at any moment during the exposure is what is illuminated by the plane of light (obviously the room has to be dark).

5. This results in every part of the frog being "equally close" to the film plane, so there is no perspective distortion. Also, since every part of the frog was passed through the plane of light, every part is in perfect focus.

6. This method obviously results in harsh lighting, as the light is coming from one, two, or three light source(s) of the plane of light. Hence, my idea for a circular light-source to be placed around the sample.

Note: There are limits to how much depth you can get with this method. Obviously, the exposure starts out with the part of the sample closest to the camera illuminated by the plane and ends with the part of the sample furthest from the camera illuminated. The sample's depth hence cannot be greater than the camera to subject distance. e.g., at 3x, a 90mm lens focuses on something 4.72in away. The sample hence cannot be more than 4.72in in depth, otherwise it will run into the lens while the motor is bringing it forward.

* Or if you tilt the film plane, you can have the plane of light not parallel to the film plane, but such that it intersects with the film plane (Scheimpflug's rule).

[Richard Mc]

have a look at the web page ->

http://www.goldstreetstudios.com.au/

That photo of the snail shell was taken using the scanning light macrophotography technique.

This nicely shows the infinite depth of field that you get with this technique, but it's not so good for objects with concavities as the planar light can't always get in to illuminate these spots.

Ellie has such a set-up and may have some insights regarding potential gotchas and pitfalls.

She is pretty devoted to film, large format and alternative processes, and her setup involved a vertically oriented Sinar Large Format Camera.

I gather it can be a bit fiddly to focus (lots of bellows extension means pretty dim image on GG requiring supplemental illumination to see anything easily), but once you have that right, then the camera and light plane stay still and the subject moves up (or down) at constant speed through the plane of light with the room otherwise in total darkness.

Of course, the object may stay still and the camera/light plane could do the moving.

The constant speed results in all parts of the object being exposed equally.

If the stepper motors are not smooth enough then you get banding caused by uneven exposure.

The main thing is that the camera is always focused exactly at the plane of light so that every part of the subject that is illuminated is in perfect focus - by sweeping the plane of light over the object, you end up with an image with all parts of the object in perfect focus.

In many ways it is like the analog version of focus stacking (eg. Helicon Focus software).

Last time I spoke to her about it, I think she was planning to decommission the unit and build a bigger and better one.

She has connections at a local University and I think the technicians in the Physics Department may have some expertise in building these light sources.

Anyway, it might be worth dropping her an email.

[dh003i]

Richard,

Thanks a lot for the link and reference!

have a look at the web page ->

http://www.goldstreetstudios.com.au/

That photo of the snail shell was taken using the scanning light macrophotography technique.

This nicely shows the infinite depth of field that you get with this technique, but it's not so good for objects with concavities as the planar light can't always get in to illuminate these spots.

Yes, this has occurred to me; it was partly why I used a frog in my example -- they don't have many cavities. But an ear -- as creepy as it might sound to put an ear on a focusing stage -- would present that kind of problem. It may be solvable by doing multiple exposures with the plane of light in different orientations, and the film or lens-plane adjusted to focus on those planes...but that would require perfect repositioning of the sample back to the zero-point, which might be difficult.

Ellie has such a set-up and may have some insights regarding potential gotchas and pitfalls.

She is pretty devoted to film, large format and alternative processes, and her setup involved a vertically oriented Sinar Large Format Camera.

I gather it can be a bit fiddly to focus (lots of bellows extension means pretty dim image on GG requiring supplemental illumination to see anything easily), but once you have that right, then the camera and light plane stay still and the subject moves up (or down) at constant speed through the plane of light with the room otherwise in total darkness.

Yea, f/4 as indicated by the aperture-ring on a symmetrical lens is effectively f/16 at 3x (or in the case of a reversed asymmetric Micro Nikkor 55/2.8 with a 0.8x pupillary factor, effectively f/17). Dim light.

Of course, the object may stay still and the camera/light plane could do the moving.

Thanks for the suggestion! This hadn't even occurred to me. It seems like this would be better for avoiding subject motion, as the subject would then be completely still. OTOH, you would then have distorted subject perspective, as the nearest part and furthest parts of the subject would be at different distances from the lens when illuminated.

The constant speed results in all parts of the object being exposed equally.

If the stepper motors are not smooth enough then you get banding caused by uneven exposure.

Any suggestions on a stepper motor and focusing stage? Kazilek and Sharp -- the authors of the article I linked to -- use a modified microscope stage and focusing motor. For focusing motors, the only thing I've been able to find is a JMI telescope focusing motor. Velmax elevating tables are $432, and that's for a manual one without a motor control.

Anyway, it might be worth dropping her an email.

Yea, I'll definitely do that!

[charleymeyer]

As a research geologist I was very interested in this technique in the 1980's when several articles were published and I even considered constructing one. There was a commercial instrument available that was ~40K. The best light source is the the projection of a thin slit, producing a band of light less thick than the depth of field of the lens. This can be accomplished by mounting a pair of razor blades in a slide mount nearly touching so the projected light slit is thinner than the depth of field. Three projectors are usually used and aligning them won't be simple. Using a diffuse light source will defeat the criteria I previously mentioned (sharply defined slit thinner than the depth of field of the lens). The commercial version had a heavy hydraulic stage that was lowered by bleeding pressure, controlling the exposure by controlling the rate of pressure release. Using a stepper motor was a possibility I considered but I never had the time or budget to pursue.

[JRFrench]

You can get devices that project a line of red laser light from many sources, and they are only a few dollars. They would be an interesting place to start. The lens assembly might be able to be modified.

http://www.dealextreme.com/details.dx/sku.5928

A stepper motor on a worm drive would be well suited to the stage, no slip and low gearing.

[Nathan Potter]

This is an interesting technique, but requires considerable precision. I rigged this up on a couple of occasions for some industrial photography. I moved the sample on a free falling Newport Inc. stage that fell by gravity using a dashpot to control the descent of the sample. The light source was three gooseneck halogen illuminators arranged around the sample at about 120 degree intervals so the sample was illuminated on all sides (not always necessary depending on the sample). I found it necessary to use two concentric peripheral slits one inside the other to obtain a clean slit of light at the sample. Obtaining a precision small slit width was problematical not to mention then aligning the point of focus exactly at the slit illumination. That alignment requires that the plane of the slit illumination be exactly parallel to the focus plane unless as mentioned above a Schiempflug condition is desired where the alignment becomes near impossible due to its becoming a three dimensional problem with the illumination at some angle to the film plane. For alignment I used a piece of clear glass with X and Y strips of reflective paper as the sample to check that the camera focus was exactly at the slit of light over the field of view.

I used slits made of shim stock folded into about 6 inch diameter rings. The slits formed with two pieces brought together and spaced precisely using, as I recall 0.010 inch temporary shims to maintain exact gap while wire supports were epoxied for permanent gap maintenence. Both inner and outer slits were made this way. Shim stock was brass so the wire supports could also be soldered I suppose.

Thus my acceptable COC was set at 0.010 inch, (250 um) for the 4X5 setup with a 150 mm enlarging lens, (I think it was an older Componon).

All pieces have to be fanatically locked down. I was lucky to have an optical bench to work on along with some quality hardware. Surprisingly the dashpot vertical motion was very uniform, showing no banding.

I optimized exposure using polaroid film which I highly recommend - well Fuji now I guess. Although one could calculate pretty accurately knowing the light intensity, downward sample velocity, and the width of the slit.

I like the comment above "whoose" which neatly explains my impression when first reading about the technique some years ago as well as my bosses impression when I tried to explain it to him.

Nate Potter, Austin TX.

[dh003i]

As a research geologist I was very interested in this technique in the 1980's when several articles were published and I even considered constructing one. There was a commercial instrument available that was ~40K. The best light source is the the projection of a thin slit, producing a band of light less thick than the depth of field of the lens. This can be accomplished by mounting a pair of razor blades in a slide mount nearly touching so the projected light slit is thinner than the depth of field. Three projectors are usually used and aligning them won't be simple. Using a diffuse light source will defeat the criteria I previously mentioned (sharply defined slit thinner than the depth of field of the lens). The commercial version had a heavy hydraulic stage that was lowered by bleeding pressure, controlling the exposure by controlling the rate of pressure release. Using a stepper motor was a possibility I considered but I never had the time or budget to pursue.

Yep, I saw that suggestion for the light-source in the article I linked to. I actually have the projectors, but would like the possibility of something else as positioning and aligning them will necessarily be a PITA. That's why the idea of using hundreds or thousands of fiber-optic cables arrayed in a 360-degree circle around a donut-shaped thing occurred to me (of course, positioning and gluing these fiber-optic cables between two donuts would be a PITA, but it'd only be done once). The many light-sources are necessarily aligned because they're sandwiched between two donuts, which keeps them planar with one-another.

As you said, the issue is eliminating the diffuse light. My idea for that was to recess the fiber-optic cable terminal ends into the donut, so that non-planar light is eliminated, like so:

Side view of donut with fiber-optic in middle:

===================
====
===================


The black lines represent the top and bottom donuts (painted black on the inside) sandwiching the fiber-optic terminus (which is the red line). My idea is that because the cable is sufficiently recessed into two black overhangs, light not traveling sufficiently planar is eliminated (absorbed into the black paint as heat).

[dh003i]

Nathan,

Thanks for your comments.

This is an interesting technique, but requires considerable precision. I rigged this up on a couple of occasions for some industrial photography. I moved the sample on a free falling Newport Inc. stage that fell by gravity using a dashpot to control the descent of the sample.

Newport Inc. seems like a company similar to Velmax, which I mentioned above (Velmax is actually in my local area). The idea of a free-falling stage hadn't occurred to me, thanks! This seems like it'd be cheaper and maybe easier than a motorized stage, while just as accurate. What should I look for from Newport's products? They have a bewildering array of products. What kind of dashpot would you recommend? I'm having some difficulty visualizing it; e.g., how something like this interfaces with the free-falling stage to control and linearize the rate of descent.

How does the dashpot eliminate acceleration upon descent? Normally, falling objects accelerate at 9.8 m/s/s, so the rate of descent initially is slower than that at the end.

The light source was three gooseneck halogen illuminators arranged around the sample at about 120 degree intervals so the sample was illuminated on all sides (not always necessary depending on the sample). I found it necessary to use two concentric peripheral slits one inside the other to obtain a clean slit of light at the sample.

So what you are doing is something like this, essentially double-masking the light-source (wouldn't that be similar to my proposed overhang method?):


===> |
===> ||||
===>
===> ||||
===> |

Obtaining a precision small slit width was problematical not to mention then aligning the point of focus exactly at the slit illumination. That alignment requires that the plane of the slit illumination be exactly parallel to the focus plane unless as mentioned above a Schiempflug condition is desired where the alignment becomes near impossible due to its becoming a three dimensional problem with the illumination at some angle to the film plane. For alignment I used a piece of clear glass with X and Y strips of reflective paper as the sample to check that the camera focus was exactly at the slit of light over the field of view.

Hmm, why not first align the light-sources at the angle that you want the sample illuminated, then use the camera's movements to position its focus plane?

I used slits made of shim stock folded into about 6 inch diameter rings. The slits formed with two pieces brought together and spaced precisely using, as I recall 0.010 inch temporary shims to maintain exact gap while wire supports were epoxied for permanent gap maintenence. Both inner and outer slits were made this way. Shim stock was brass so the wire supports could also be soldered I suppose.

This is a little bit difficult to visualize. Are you saying you made two concentric cylindrical shaped slits positioned around the sample, and then put the goose-necks behind the concentric slits, like so:

Attachment 44556

Might this make an acceptable shim material?

Oriented Polyester Shim Stock, Silver, 0.0005" Thick, 5" Width, 20" Length (Pack of 1)

Thus my acceptable COC was set at 0.010 inch, (250 um) for the 4X5 setup with a 150 mm enlarging lens, (I think it was an older Componon).

Hmm, a 250 micro-meter CoC? That's 0.25mm, which wouldn't even seem to support enough resolution for contact-printing (0.2mm acceptable CoC for 25cm viewing of final print / 0.25 = 0.8x enlargement). Am I missing something?

All pieces have to be fanatically locked down. I was lucky to have an optical bench to work on along with some quality hardware. Surprisingly the dashpot vertical motion was very uniform, showing no banding.

I have a Bencher M-3 Illuma Copystand with a counterbalance to support the 4x5 monorail. It looks similar to this when reassembled (I disassembled it to clean it), but without the accessory lights. What else will I need to make sure that everything remains motionless?

I optimized exposure using polaroid film which I highly recommend - well Fuji now I guess. Although one could calculate pretty accurately knowing the light intensity, downward sample velocity, and the width of the slit.

I'll probably try it out with Kodak E100VS first, as I have tons of that film. Any references on calculating the exposure?