After completing my first optical enclosure and a follow-on hastily-built "cheap" enclosure, I decided that it was time to build yet another high-performance optical transceiver assembly. I had onhand a pair of larger Fresnel lenses, approximately 404mm x 430mm (about 16"x17") and their focal length was approximately 229mm, or 9 inches. These Fresnels were purchased from Surplus Shed and were part number L3707. Note: As of 6/07, it appeared as though these lenses were out of stock.
 |
 |
Rather than have a single unit that held both lenses, I chose to build two completely independent lens enclosures, and when completed, these would be connected together with hinges, allowing the unit to fold up for easier transportation. It was most important that, during operation - when unfolded - that their mechanical alignment not only had to be be stable, but repeatable.
The main enclosure bodies:
As with the first enclosure, both enclosure bodies were constructed using less than a single 4' x 8' sheet of "5.2mm Hardwood Plywood" (the actual thickness of which is about 4.6mm) which was obtained at Lowes for about $12. These sheets are 3-ply (not counting the two very thin exterior veneers) and the "finished" veneer (being "A" grade) being thin enough that it cannot take much sanding at all, while the obverse veneer is somewhat thicker and is of "C" grade. It is worth mentioning that sheets of plywood this large and thin are not particularly flat - something to be considered during cutting and construction as this uneven-ness needs to be accommodated during construction.
Because the material is so thin, there is relatively little edge surface into which nails may be driven or glue applied. To solve this problem - plus add extra strength overall - scraps of this plywood (from the original rough-cutting of the larger pieces) were used to reinforce the edges: These were glued and stapled into place (with the staples never having been removed) to assure a tight bond. After the glue was dry, a plane was used to true the edges to the appropriate angle for later assembly.
 |
 |
 |
 |
Later, the pyramidal structures were assembled very carefully. After checking angles - and planing the mating surface as necessary for best contact and fit - the pyramids were roughly assembled using masking tape. The fit was checked yet again and then tacked together using very small steel brads. After yet another check of proper fit and alignment, the nailed-together joints were pried slightly apart and glue was forced into the gap. Closing the gap again and using a spring punch to seat the brads, the excess glue was removed and the pyramids were placed, wide end downwards, on a hard, flat surface (my kitchen table) and some weight (a few small lead-acid batteries) was placed on the small end to force the front to be flat and true: Necessary adjustments were made in order to make sure that the pyramids were also square.
After the glue dried, the pyramids were much stronger - but still fairly flimsy. In order to solve this problem - and to make attachment of the pyramids to the lens frame easier, another rectangular frame was made using some 1"x1" square poplar dowels. One complication of this frame piece was that its inside surface had to have an angle that matched the outside angle of the pyramid in order to maximize the bondable surface area - yet it had to fit snugly at the very end of the pyramid. After some very careful measuring and calculating, I set up a a table saw to cut the frame pieces at the appropriate angle. Using each pyramid as a reference, I then cut each piece to length, dry-fitting as I went. The final step was to assemble this frame piece, using glue, long screws, and (an absolute necessity!) four corner clamps to hold the frame square during the drilling of pilot holes for the screws as well as their installation, and while the glue dried. As with other pieces, this frame was weighted down onto a table to maintain flatness while the glue dried.
After the glue on the 1"x1" dowel frame had dried, it was fitted over the pyramid shape at the front. Using lots of clamps and a few temporary brads, the frame was loosely tacked into place, the fit and alignment checked, and then wood glue liberally applied - along with more brads. After cleaning up the excess glue, the pyramid was again weighted down - again, with the front against a flat surface while the glue dried. After drying, the pyramid was quite rugged and I was ready for the next step.
On the rear of the pyramid (the small side) a plywood plate was attached. This plate, made of a double-layer of plywood (two pieces glued together and clamped between aluminum plates) was glued and tacked into place, completing the main pyramidal structure. After the glue had dried, a hole saw was used to cut an opening for the mounting of the electronics. At this point, the pyramid structure was more than strong enough to easily support 220 pounds (100 kg) without the slightest fear of damage.
Now, all brads were set below the surface of the wood and wood filler was applied to the holes, as well as to other voids. After the filler had dried, it and the wood were sanded carefully (to avoid breaking through the thin veneer) using a fine-grit paper and it was time to apply finish to the wooden pieces.
 |
 |
 |
 |
 |
 |
 |
 |
As a final step, the inside surface of the pyramid was sprayed with flat black paint - a precaution done to minimize the response of the receiver to off-axis, stray light sources.
With everything having been stained, painted, and/or varnished, it was now time to begin final assembly of the unit.
The main purpose of 1"x1" poplar frame around the edge of the pyramid is to provide good attachment points to the rest of the assembly. Pilot holes were drilled in this rail and, using drywall screws, the pyramid was first aligned and then attached to the lens rail, pieces marked for later reassembly, and then removed. At this point, three "bi-fold door" hinges were used to attach the two lens frames to each other: Three hinges were used for for maximum stability, and these types of hinges were used because of their ability to fit flush between to pieces of wood as well as avoid interference with themselves and other pieces as they were opened and closed.
Hinges and truss assemblies:
When deployed, the two lenses are side-by-side and in the same plane - one of the jobs of the hinges. In order to keep the transmit and receive beams parallel, however, it is necessary to make the entire structure rigid when assembled, preventing any misalignment. To achieve this, a wooden truss assembly was constructed, bolted into place when the unit was unfolded. In order to provide attachment points for this truss assembly, some steel "tee" nuts were installed in the 1"x1" frame around the pyramid: When the pyramid assembly is screwed to the lens frame, these tee nuts are "captured" between the pyramid assembly's frame and the lens frame.
There are two identical triangular truss assemblies, each constructed from 1"x1" poplar and using a combination of long screws and wood glue for strength, making for extremely rigid and stable structures: Pyramidal trusses were chosen because their dimensional stability is based primarily on compression and tension of the wooden pieces rather than their rigidity and ability to stay strait. When deployed, these trusses - along with the hinges - are used to hold the assembly "open" and keep the lenses parallel to each other. For assembly, 10-24 screws (matching the tee nuts) with wing nut heads are used to bolt the truss assemblies to the pyramid structures. To prevent damage to the wood, some 1/4-20 tee nuts were installed in the truss assemblies: While ordinary washers on the 10-24 screws would have also worked, these could easily get lost - especially when trying to assemble the unit while in the dark. The tee nuts have the advantage that they are held into place, both with their built-in spikes and with a composite glue.
Elevation adjustment mechanism:
One necessary feature of any system such as this is a means by which aiming can be accomplished. The horizontal adjustment of the beam is quite easy: Simply rotate the entire transceiver. Adjusting the elevation is a bit trickier and a stable and reasonably fine adjustment is desired.
To accomplish this with this system, it was necessary to add another truss in order to provide a stable support for the elevation adjustment. This piece bolts to the pyramidal frame in the middle and to each truss on the top and bottom, anchoring the top and bottom trusses firmly: Without this extra, center truss member, the top and bottom trusses are unstable, as they are not supported in such a way that they can tolerate any vertical stress. In order to minimize space, each piece of this truss assembly is attached to another using hinges, allowing this piece to be folded up.
On this bolt-on truss assembly are two other pieces, also hinged, that attach to a screw assembly. On the bottom piece, there is a sort of "kickstand" that, as the screw assembly is adjusted, pushes down, causing the lens to tilt downwards. One minor complication that the entire transceiver is somewhat "front heavy" and can be blown over in a high wind. To prevent this, the lead-acid batteries used to power the electronics are simply placed atop the bottom truss to weight it down.
Electronic assembly mounts:
Around the holes in the rear mounting plates are four 10-24 tee nuts used to hold the transmit and receive electronics into place: The electronics unit is simply installed in the hole and four screws are used to hold it firmly - and accurately - in position.
The electronics packages themselves are mounted on plates made from 0.062" glass-epoxy circuit board. Mounted to this board, on the front side, is a thin ring of 3" ABS wastewater pipe as seen in Figure 4 - the purpose of which is to assure proper alignment of the optics as well as to simplify installation: As it happens, the size of the hole cut in the back of the pyramid is a snug fit with the ABS pipe section, providing good physical alignment for the optics.
Initial lens focusing and alignment:
Once the unit was assembled, I followed the same procedure used in the initial alignment of the first enclosure - that is, a Laser level was attached to a carpenter's square and used to point the plane of the lens at a paper target located across the basement. At this point, the receiver and transmitters were installed and roughly aligned to be centered on the points marked on the target. For final alignment the same procedure was followed for this transceiver as was followed for fine alignment of the first transceiver - that is, with a more-distant target and "optical beacon" system.
 |
 |
 |
 |
 |
 |
Comments about the use of short focal-length Fresnel lenses:
You may have noticed that the Fresnel lenses used have a very short focal length - on the order of 0.6 or so. When the Fresnel lenses were ordered, Surplus Shed's catalog initially listed them as having a 24" focal length, but when they arrived, their focal lengths were measured to be just 9". Shortly after this, Surplus Shed must have realized their error and amended their catalog, listing the two lenses separately, having realized that they had a mixture of both 9" focal length and 24" focal length Fresnels of the same dimension. On advantage of the use of 9" focal length lenses was that, potentially, a much smaller enclosure could be used.
Soon after this, I made contact with Chris, VK3AML, who advised that such a short focal length could be a liability, owing the the difficulty in properly illuminating the optics and getting good far-field luminous intensity as well as potential problems with receiver sensitivity. Suffice it to say, I was expecting to encounter some challenges when interfacing these lenses with the electronics.
Initially, it looked as though the Lambertian pattern of the Luxeon III was a reasonable match for the apparent angle of the lens from the vantage point of the focal point without the use of a secondary lens. After constructing the enclosure and doing surface luminosity measurements, it soon became apparent that secondary optics would, in fact, be required: Not only was much light "wasted" going past the edges of the box, but the amount of light reaching the Fresnel dropped off rapidly from the center of the lens, toward the edges.
After trying an assortment of PCX (Plano-ConveX) and DCX (Double-ConveX) lenses, I immediately noticed a problem: While I could better-concentrate the light into the Fresnel - something that required a fairly low-diopter lens - there was still the problem with edge-falloff. The main problem was that at the edges, the distance between the Fresnel and LED was about 1.5 times that of the distance from the center of the lens to the LED - a property that explained the rather dramatic dropoff.
In order to remedy this problem, an aspheric lens was required. Fortunately, I had onhand a quite a few inexpensive, optical-acrylic DCX lenses to experiment with. After a few hours of experimenting, I was able to regrind one of these DCX lenses such that it fairly evenly illuminated most of the Fresnel lens. At some point, I may do more testing and further refine this type of lens, as there is some luminous dropoff at the extreme edges of the Fresnel.
Comment: Acrylic lenses are easy to re-grind. I simply used a file to set the appropriate, rough lens shape, then switched to progressively finer sandpaper (150-300-600 grit) with final polishing done using Brasso (tm) polishing compound. As can be seen from the upper-right picture in Figure 4, a good optical result can be obtained with reasonable care.
How does this transceiver compare in performance to the first? Because of the "custom" secondary lens, the far-field optical flux of this transceiver is over 20% greater than the "first" transceiver - a result of more efficient illumination of the Fresnel lens.
Should I use a short focal-length Fresnel lens, then?
The upshot? While it is possible to use such a short focal-length lens successfully, the complications are such that I would recommend that one use a lens with a focal length of 1.0 or greater unless you are
prepared to "jump through hoops" to make it all work! With longer focal lengths, the complications mentioned above (e.g. having to grind custom lenses) are minimized.
One question that has come up has to do with the quality of short and long focal length Fresnel lenses. If you get good-quality Fresnel lenses, of the size of the blur circle is more or less proportional to the focal length: In other words, the longer the focal length, the larger the blur circle that results. Keep in mind that this applies only to good quality Fresnel lenses that were intended to be used as good-quality image magnifiers: Other types of lenses - such as those used for overhead projectors - aren't typically manufactured to a tolerance or quality that allows it to be used to produce a precise beam.
In actual in-field testing, I have found that the beamwidth of this short focal length lens is precisely the same as that of a lens with a longer (about 1.3) focal length lens - about 0.27 degrees.
Comments:
This enclosure has been transported several times and seen several in-field trials. It has been found that with the use of the rigid truss assemblies and the pyramidal construction that the paraxial alignment of the transmitter and receivers has been precisely maintained.
Return to the KA7OEI Optical Communications Index page.
Copyright 2007-2008, KA7OEI.
Last updated 20081023