What's new
What's new

light sources for autocollimators & alignment scopes

rimcanyon

Diamond
Joined
Sep 28, 2002
Location
Salinas, CA USA
I have managed to acquire a couple of optical alignment tools that came without light sources. One is a Leitz autocollimator that has a fiber optic cable protruding from the body of the scope, but no source of light. The other is a K&E alignment telescope, also with no source of light (just a socket with a mirrored prism at the bottom). I would appreciate some help finding the right parts to get these working, since my web searching has not turned up much except new parts, and I don't know a lot about light sources, like the wattage and type of bulb needed.

-Dave
 
Dave --

Most K+E line-of-sight and alignment telescopes "need" a special lamp housing that plugs into the telescope's autocollimation adapter. The lamp inside the housing is a "T 1 3/4", and K+E supplied #327 lamps that operate at 6 volts.

On these instruments K+E relied on the centering of the filament within the lamp to get the bright spot into the working area of the reticle, no lateral adjustment is necessary. This makes the K+E design improvisation-friendly . . . just about any light source that will fit into the telescope adapter will work. I've even used standard flashlights by simply shining the beam into the lamp-socket hole in the adapter.

Let me suggest you consider making your own light source using a high-intensity LED instead of an incandescent lamp. At the direct-but-kludgy end, you could simply make a temporary bushing to hold a tubular LED in the center of the socket and use a rubber band or tape to hold the bushing and LED in the telescope, or you could make something more elegant.

Incidentally, autocollimation adapters necessarily contain some sort of a beamsplitter that diverts a portion of the light coming through the telescope. For this reason, telescopes with autocollimation adapters installed are "dimmer" than similar telescopes without the autocollimation adapters. If you need to use the telescope as an alignment telescope (ie, to view finite objects rather than autocollimation images reflected from a mirror) in low-light conditions, you can increase the brightness of the image you see in the telescope by removing the autocollimation adapter.

If you have the K+E autocollimation adapter I think you have, you can remove the adapter by removing the two screws holding the adapter into the side of the telescope eyepiece barrel. You will want to cover the hole in the barrel, and K+E made a sheetmetal cover than screwed in place. But as it happens, a 5/8-inch-or-so band cut from a Kodak 35 millimeter film cannister will slide over the eyepiece barrel with just enough friction to hold it in place.

I can't help much with the Leitz autocollimator, but most fiber-optic illuminators use a very-high-intensity source that are pretty-much interchangeable.

John
 
I don't know how important it is to gather all original equipment or how detailed you wish to make your restoration but, if it's only light you need I suggest using the white LED's that are now available. They are powerful and efficient emmitters of almost white (a little bluish-purple actually) for their size.

I've seen figures where white LED's are about 4 times as efficient as flourescents in terms of light output for power consumed which are about 3 times as effeicent as incandescent. Radio shack now has them and there's several sources under "White LED" in Google.com.
 
Autocollimators are not too critical in type light source nor do they need to be particularly bright. My Leitz autocollimator has a built in green filter so a green LED would work well. So would the white LEDs as Forrest mentioned. I machined an aluminum barrel to fit a AA battery Minimag flashlight to my autocollimator. Works very well and self-contained...no external power supply.

Don Clement
Running Springs, California
http://www.ClementFocuser.com
 
John, the K&E fitting looks like this:
ke.bmp


The K&E is on the way from Seattle. I don't have it yet, so I am trying to get ready to start using it. It includes an optical square attachment, and it has a spherical mount that slides onto the scope barrel. I have some mirrors already, as well as a couple of round targets that I can mount as needed and a collimating light source (also in need of a light fixture).

I would like to set up my lathe with a pair of targets at either end of the spindle bore, align the scope to that axis, then verify headstock alignment with the bed, cross slide perpendicularity, bed wear, tailstock height and tailstock ram alignment. Having two scopes may not be needed but I was not sure which kind of scope and reticle I would find most useful.

Reticles. Do they all need to match, or can I do a good job of alignment without matching reticles? The K&E comes with the reticle below:

ke2.bmp


the Leitz reticle has a linear scale that runs across the field of view and which provides direct reading 0-30 (I don't know what the units are; seconds perhaps?). The optical assembly can be rotated 90 degrees for measurements in both directions. It also has a peripheral scale that runs around the outside edge of the field of view, and is marked 0-60 and can be rotated independently . (I don't know what the units mean, the 0-60 scale consumes about 270 degrees of the field). The eyepiece can focus on the reticle or on objects from about 30 feet to infinity, but not in between. Maybe this is ok with an autocollimator, even when the targets are much closer, the reticle is what is important???

The targets and the collimating light source have a third type of reticle, with three pairs of lines that run through the center of the field.

The main things I am still missing (besides experience and knowledge in this area) are adjustable mounts for the scope and for the collimating light source, if I end up using it. If any of you have made your own, I'd like to hear the details. I've seen pictures of mounts with two pairs of rotating cones that look like they can be fabricated, but some close-ups and details would be useful.

Dave
 
Auto collimators need a more or less mono chromatic light source. Usually green. A white LED will produce significant dispersion and therefor inaccuracy in the optics. A green bright LED should work well. The eye is most sensitive to green light.
 
The basic autocollimator consists of a lens, reticle, beamsplitter, and measuring scale. Some autocollimators (Leitz) increase accuracy by using a tilting plane parallel glass at the scale end. Dispersion has some effect on accuracy if the objective is a single glass lens and/ or the beamsplitter is not the pellicle type. A BPF filter should minimize the effects of dispersion. Green is typically used because the eye is most sensitive to green.

My current project is building a dual mode electronic autocollimator /alignment scope. The measuring scale and plane parallel glass are replaced with a United Detector four quadrant silicon diode detector that I have. The two outputs from current to voltage difference op amps circuit are feed to A/D PC card. Software is “wired” using NI LabView that will display XY angle deviation in autocollimator mode or XY displacement of the centroid of a laser in alignment mode.


Don Clement
Running Springs, California
 
Dave, its interesting that your trying to set up a autocollimator for lathe alignment. I have two of the old style Hilger & Watts units to do the same thing. The problem I encountered was finding the proper targets. Unfortunately, the mirrors need to be very accurate and the mount square. I found a couple on ebay, but they went for big bucks. I guess
thats the trick. Let me know how you are going to set this up. We have since made a test bar and are doing it the tied and true method. But I wish I could use the autocollimators
 
Dave --

Your pictures are of a K+E Alignment Telescope, which may have an autocollimation attachment. A direct descendent of this telescope is still being made by Brunson Instrument Company of Kansas City, Mo. (Brunson purchased the remains of K+E's Optical Tooling business and product line about a half-dozen years ago.)

This type of instrument designed to have a very straight "line of sight" (LOS) through its entire focus range that is coaxial with the barrel of the telescope. If I recall correctly, the Aircraft Industries Association specification for the telescope barrel is that the diameter must be between 2.2495 inch and 2.2498 inch with no more than 0.0001 inch of taper.

The "spherical adapter" is a 3.5000 inch nominal diameter sphere with a 2.250 inch bore through its center. When the adapter is mounted on the telescope barrel, the should be within 0.001 inch of the sphere center.

The sphere is designed to nest in a female-conic mount (often referred to as a "candlestick" that is usually attached at at both ends an assembly jig or fixture such as those used to build large aircraft.

The centers of 3.5000 inch diameter spheres -- one holding an alignment telescope and the other holding an optical target -- physically define a virtual reference line. Pointing the telescope at the target puts the telescope LOS on the virtual reference line, allowing measurements to be made relative to the virtual line.

So the next question becomes how to point the telescope toward the optical target. Here's where more hardware is needed: an adjustable mount to hold the other end of the telescope and allow for precision pointing.

Speaking kinematically, the sphere controls three degrees of freedom of the telescope-and-sphere assembly and we don't care greatly about the rotation of the telescope LOS around the virtual line. This means that the adjustable mount needs to control two degrees of freedom.

V-block time. The commercial mounts that I'm familiar with are essentially solid V-blocks that can be adjusted in two perpendicular directs that are both perpendicular to the reference line OR they are some V-block variant that allows the two sides of the V to be adjusted independently. The cone mount you speak of does just that; the conical whole angle is 90 degrees, and two of these cones are mounted adjacent to each other on threaded shafts so that the adjacent sides form a V in which the telescope barrel nests.

Adjusting the "height" of one cone obviously raises or lowers the whole cone, but it effectively moves one side of the telescope-mount V up-and-right/down-and-left or up-and-left/down-and-right, depending on which cone is adjusted.

The telescope itself is adjustable for focus (the knob attached to the satin-chrome micrometer collar) and has two "optical micrometers" (the knobs attached to the white micrometer collars) that displace the LOS laterally up to 0.050 inch in opposite directions. The optical micrometers are arranged to operate in perpendicular directions. These directions are often described as X and Y directions, but I'd STRONGLY suggest using the more-rigorous terminology of TELESCOPE X and TELESCOPE Y to positively distinguish between Telescope coordinates and Object coordinates.

I've just been clobbered with a hint, so I need to cut this short and I'll be out of town tomorrow. So let me wrap up with a couple of quick points.

What you term the reticle isn't. It is an auto-reflection target on the objective lens cover. The real reticle is inside the telescope down at the eyepiece end. You can see the reticle if you unscrew the eyepiece, it looks like a small disc of bluish glass in a brass ring. The reticle pattern is scribed or etched on the glass, but the line width is only about 0.0002 inch so it's hard to see, and it's delicate. If you're lucky, your instrument will have a "filar / bifilar" reticle, but most of the K+E Alignment Telescopes with the dark-green micrometer housings have single-line "crosshair" reticles.

Let me suggest that you spend some time on Brunson's website; they have an optical-tooling tutorial section that is excellent although it tends to concentrate more on jig transits than alignment telescopes. Nevertheless, the principals are the same and you'll find the information useful.

If you have any specific questions, do post back and I'll try to answer.

John
 
John, thank you for the informative reply. The Brunson tutorial was pretty good, at least it helps get the terminology down. I am considering purchasing the Boeing optical alignment manual that has been copied and sold on ebay for additional information.

I still have a couple of questions about the Leitz autocollimator:

The Leitz autocollimator eyepiece will focus on the reticle fine, but it will not focus on anything between 0 and about 30 feet. It will focus on something that is 200 feet away and the depth of field is relatively small at that depth, i.e. maybe 30 feet. Can it still be used for short distance autocollimation, where short is less than 10 feet? i.e. will the image of the reticle that is reflected back be in focus even if the eyepiece is not set for 2Z focus (Z being the distance to target)?

The outer scale that rotates around the field of view: what is it for, and what do the numbers mean (0-80 would be a 360 degree rotation)?

Dave
 
Dave --

Autocollimation means "self collimation", and requires reflecting a collimated beam projected from a telescope from a very flat mirror so that the reflected beam goes back into the telescope. The instrument's objective-lens system must be focused to "optical infinity" to autocollimate, and almost by definition the objective lens system of an instrument called an "autocollimator" would not focus to "near" or "finite" targets.

An optical tooling instrument intended to focus on finite targets would be considered a telescope, and most optical tooling telescopes can be equipped to autocollimate when focused to optical infinity.

In an optical sense, such telescopes would have an objective lens system (which includes the focusing lens or lenses), a reticle, and an eyepiece (aka "ocular" lens system). The job of the objective lens system is to create a "real image" of the target at the reticle plane, while the reticle is in essence simply an index marker. The ocular lens system is a high-power magnifier or microscope to magnify the real image and reticle so that the eye can discern details.

There will be two different focus adjustments for an optical tooling telescope, one to focus the objective lens system on the reticle plane, and a second to focus the eyepiece on the reticle plane containing not only the reticle itself, but also the real image created by the objective system.

On your K+E Alignment Telescope the objective focus knob is the one with the metallic collar, while the ocular focus is controlled by a knurled drum surrounding the outer eyepiece lens. (The ocular system is threaded into a fixed bushing, and the whole ocular system moves toward or away from the reticle.)

Your Leitz autocollimator would also have an ocular focus adjustment, and it sounds like you have been using the ocular focus to try to focus the telescope.

Standard practice when focusing the ocular would be to point the instrument toward a light field -- I usually hold a white index card in front of the objective to reflect ambient light -- and adjust the eyepiece until the reticle is sharp. This ocular focusing usually has to be done incrementally, as the eye automaticaly adjusts to compensate for residual ocular focus error in just a few seconds, but the eye fatigues readily if not relaxed.

So after a few seconds you will want to look away from the eyepiece and not look at anything in particular to let your eye return to its relaxed state. Then quickly glance into the eyepiece and tweak the ocular focus as necessary. Several iterations of the focus-relax-focus cycle may be necessary, especially when you are first learning how to focus the eyepiece.

Only after the ocular lens system is properly focused on the reticle is it possible to properly focus the objective system.

The objective system needs to be focused to the reticle plane, not just until the target image appears to be sharp. If the objective is nearly focused, the eye will adjust to optimize the focus of both the reticle and and target image, which leads to eye fatigue. But there's another problem with a nearly-focused objective that contaminates the measurements you are trying to make . . . parallax.

Parallax is the visual change in position of one object relative to another caused by a change in observer's location, and in this context it is what causes the reticle seem to move relative to the target if the observer's eye is raised, lowered, or moved side-to-side. And the "secret" to eliminating parallax?? Move your head up-and-down or right-and-left just a bit while adjusting the objective focus; when the reticle appears stationary on the target the parallax is at its minimum.

I'm running out of lunch hour and need to run for now.

John
 
John,

I have looked for another focus adjustment without success. I may need to find someone with a Leitz autocollimator to explain. The tube that surrounds the objective unscrews, and there is a large O-ring that it compresses against the lens tube as a dust seal. The other knobs do one of three things: adjust the reticle up/down, rotate the reticle and eypiece assembly 90 degrees, and rotate the reticle that has the peripheral scale on it. Beyond that, the body of the scope is plain, with two access hatches held on by screws.

Thank you again for your detailed explanation; I have learned a lot; time to acquire some field experience.

Dave
 
Not exactly on topic, but since we seem to have attacted a very knowledgeable party, here goes.

Is it practical/effective to use an old K&E jig transit with the micrometer on the objective end in combination with scale type targets as an aid to judge scraping progress on a long lathe bed?

John
 
To Dave --

Autocollimation requires that the autocollimator's reticle be at the focus of the objective lens system . . . no focusing is needed beyond the initial adjustment of the instrument at its assembly.

The standard "target" for an autocollimator is a plane mirror with a surface flat to within 1/4 of the wavelength of the light of the autocollimator beam. The distance between the mirror and the autocollimator is immaterial to the autocollimation measurement; the mirror could be a hairsbreadth away from the objective or across an auditorium without creating any need to refocus the autocollimator objective.

Autocollimators are very sensitive to out-of-plane rotations of the mirror, but they are totally INsensitive to in-plane mirror rotations and are also INsensitive to all translations of the mirror.

Although autocollimators are typically used to measure deviations from "straight", they can only do that if the non-straightness induces a out-of-plane rotation of the mirror. This is usually done by mounting the mirror on a three-footed "sled" in such a way that the line connecting two of the sled's feet are parallel to the plane of the mirror while the line connecting another pair of the sled's feet are perpendicular to the plane of the mirror.

The mirror-and-sled assembly is fundamentally a sine bar; the distance between the perpendicular-to-the-mirror-plane pair of feet is the conceptual equivalent to the distance between the rolls of a standard sine bar. Raising or lowering one of these two feet relative to the other tilts the sled and mirror, and the amount of tilt can be measured with the autocollimator.

Let's say for discussion purposes that the sled feet are 5 inches apart and that the autocollimator shows that the sled has tilted 5 arcseconds. The rise-or-fall of one foot relative to the other is calculated this way:

( Rise or Fall) = (foot-to-foot distance) x Sine (Angle of Tilt)

(Rise or Fall) = 5 inch x Sine (5 arcseconds)

(Rise or Fall) = 5 inch x 0.000024

(Rise or Fall) = 0.00012 inch

John


To John --

Back in the 1940's or maybe into the 1950's the world's authority on Optical Tooling was Dr. Philip Kissam, a Professor of Civil Engineering at Princeton University. In one of his studies, Dr. Kissam showed that with well-designed filar / bi-filar targeting the two-sigma pointing uncertainty of a high-quality commercial optical tooling telescope could be as small as 0.5 arcsecond. (Although Dr. Kissam worked with the optical tooling industry as a whole, it appears that he was more closely associated with K+E than with the other instrument makers . . . perhaps somewhat due to the physical proximity of K+E's manufacturing plant in New York City and Morristown, NJ to Princeton, but no doubt also influenced by K+E's status as the "800 pound gorilla" of US survey equipment makers.)

Dr. Kissam also demonstrated that Optical Tooling measurements, which are almost always taken as across-LOS displacements using a parallel-plate micrometer, have their own translational noise floor that is independent of the telescope pointing uncertainty.

Unfortunately my copy of Dr. Kissam's textbook "Optical Tooling" grew legs and ran out of my office several years ago and I simply don't remember what he claimed the translational measurement noise floor of parallel-plate micrometer measurements to be. If I had to guess, I'd guess 0.0005 inch under ideal conditions. (I'll note that K+E added a vernier scale to their later jig-transit-type parallel-plate micrometers to subdivide the main-scale 0.001 inch increments but their literature that mentioned the vernier almost always made a point of averaging multiple measurements.)

The US's last surviving maker of Optical Tooling instruments, Brunson Instrument Company, claims in their "Fundamentals of Optical Tooling" class that the measurement uncertainty of an Optical Tooling measurement is the greater of 0.001 inch or 1 second of arc under ideal conditions.

Since one arcsecond subtends 0.001 inch at 206+ inch, the Brunson claim can be considered to claim 0.001 inch measurement uncertainty for sight lengths up to 17 feet.

My experience -- I started doing industrial surveying in 1975 -- is that a more realistic uncertainty value for measurements performed under good "normal" conditions would be 0.001 inch per 10 foot length of sight.

Bottom line, I think that in the hands of a good instrument operator your jig transit with tooling scales and parallel-plate micrometer could be useful when roughing-in a fairly long lathe bed . . . but you'll want to add an autocollimation attachment and mirror for the fine work.

John
 
John, thank you for taking the time to share your extensive knowledge of optical alignment, and to help a novice user. I've learned a lot; I will have fun experimenting.

Dave
 
Don --

I used the term Optical Tooling as it was used by the pioneers in the field to mean a class or type of measuring equipment developed by modifying land-survey equipment for industrial applications.

Davidson has over the years "stuck their little toe" into the Optical Tooling equipment business, but I'd be flat-out astonished to learn that Davidson's management ever thought of Davidson as primarily a maker of Optical Tooling equipment.

Within my experience as an industrial surveyor, Davidson equipment is fairly common in the labs but very rare on the production floor . . . which is where Optical Tooling is done. Out on the floor, Brunson and K+E dominated the field, with a smattering of Watts, Taylor-Hobson, Gurley, Berger, and Farrand equipment to be seen.

John
 
John,

Davidson is still in business over in West Covina (near my alma mater Cow Poly) and producing optical tooling including autocollimators, alignment scopes, precision flats, etc. Davidson equipment is very common in aerospace companies here in South California.
I believe most of the pioneering work for optical tooling was done at England's NPL see K. J. Hume "Engineering Metrology" 1951
-Don
 








 
Back
Top