One of the problems with research in computational holography is finding a device to display the holograms. Unfortunately, there aren't reasonably priced commerical holographic displays. A search of the web finds commerical holographic printers which are used to print the small holograms that aqppear on credit cards. These printers are quite expensive, since there isn't a large demand for them The search will also find a number of devices that claim to holographic, but are not. It has become common practice to use holographic to describe any type of 3D display.
An interesting non-holographic technology is holographic projection screens. These are clear screens that you can see through, but also serve as projection screens. An image projected onto the screen appears to float in mid air the same way that a hologram does. These screen are relatively inexpensive and are used in marketing and advertising.
Since commerical holographic displays are not commercially available the only alternative is to make one yourself. This project started with the goal of making a holographic display for under $1000. I was willing to risk $1000 on this project and the resulting display is a good starting point for research on both the hardware and software sides of holography. The device that we've constructed is shown below along with a hologram displayed on it.
This page describes how this holographic display can be constructed for readily available parts. It also highlights some of the pitfalls in the design and construction of holographic displays.
The most important part of the design is the device that is used to display the interference pattern. For the other parts of the device there are many alternative, but that's not the case with this part. To get a quality hologram the details of the interference pattern need to be on the same scale as the wavelength of light. This means the pixels on the display should be less than 10 μm. This is a rough guide and you may be able to get away with larger pixels. Another importand display property is pixel fill. Pixel fill is the percentage of the area for a pixel that is actually covered by the pixel. A high pixel fill has a small gap between adjacent pixels. A pixel fill of at least 90% is a good rule of thumb.
With these two constraints in mind, what are the available display technologies? LCD displays clealy won't work since their pixels are too large and the pixel fill is not high enough. There are two display technologies that meet our constraints:
From a purely technical point of view LCOS is the superior technology, but it is way too expensive. While in large quantities LCOS chips are relatively cheap, small quantities and evaluation board are expensive, start at about $4000. This leaves us with DMD, which is more reasonably priced. DMD is a Texas Instruments technology that is based on representing each pixel by a small mirror. This mirror can be tilted towards or away from the user to produce a binary pattern. Since the mirrors are very small they have little mass and repond very quickly. They can typically do at least 4000 cycles/second. Different pixel intensities are achieved by modifying the duty cycle of the mirror, essentially using a time based multiplexing technique.
Texas Instruments produces a wide range of evaluation kits for their DMD devices. The TI LightCrafter is their low end evaluation board which sells for approximately $600US. This is the board that we use in our display. All the other evaluation boards in their product line have higher resolution (and higher prices), so there is a path for improving the display.
Both DMD and LCOS chips are used in projectors, do it is tempting to use a low cost project as the basis for our design. This will not work. It is very tempting to get a low cost pico projector, remove the optics and then using the display. Unfortunately, all of these devices process the image in a way that destroys the inteference pattern. Quite often the video resolutions accepted by these projectors is different from the resolutions of the DMD or LCOS chips, so the images are scaled. In addition, most of these projectors enhance to image so it looks better when it is projected.
It should be possible to harvest the DMD or LCOS chip from the projector, but they all the electronics required to drive the chip would need to be designed and constructed. It might be possible to reverse engineer the projector to extract just the parts that we need, but this sounds like too much work.
The LightCrafter right out of the box is shown below. The evaluation board is intended for projector applications so there is an optics unit on the left side of the evaluation board. This optics unit must be removed in order to access the DMD. This is the most delicate part of the process. Check that the evaluation board works correctly before you remove the optics unit, it will be hard to test it later. This is a good time to install the software and try it out.
The optics unit is attached to the rest of the evaluation board with a number of very small screws. You need to have a phillips size 00 screw driver to remove them. You can get these screw drivers at a large hardware store or a hobby store. They usually come is sets for 6 or more screw drivers, you may find the other screw drivers useful. The following picture should give you an idea of the size of these screws.
On the unit that I got the screws were very tight and hard to remove. Take your time and press down firmly on the screw. The screw heads tend to be fragile, so as soon as you loose the grip on the screw stop. In one case I had to use a Dremel to drill out one of the screws.
Once all the screws have been removed gently pull the optics unit off of the evaluation board. This is basically an upwards and to the side motion. The DMD is attached to the optics unit with a small carrier board and plugged into a socket on the evaluation board. You need to be careful to pull the optics unit horizonally away from the evaluation board before you lift it up. There are two screws on the carrier board that can be removed to free the DMD chip.
If you don't want to go through the process of removing the optics you can now purchase the evaluation kit
without the optics from Keynote Photonics (https://keynotephotonics.com/).
It is the LC3000, which can be difficult to find on their website.
I would suggest purchasing the flex cables with it, which make mounting the DMD in a
custom configuration must easier.>
There isn't much of a cost savings involved, but it does save some work.
The LC3000 looking from the top with the flex cable for the DMD is shown below.
We need a source of coherent light, which means we need to get a laser. There are many types of lasers and people doing film based holography tend to use high powered ones. We don't want to use these lasers since they will damage our DMD. We want a relatively low powered laser, the lowest we can get and still have the hologram visible. The lower power also reduces the risk of damaging our eyes. Even with a low powered laser do not directly look into the laser. I've found that a 1mW laser works well in a dimly lit room and a 5mW laser is more than adequate for normal office lighting. I wouldn't recommend using a laser stronger than 5mW.
Semiconductor lasers are relatively cheap, easy to work with and have lower power ratings. A good quality red semiconductor laser costs around $30, but you can get them for a little as $2 on some websites. I would be careful with the cheaper lasers there beam tends to be quite irregular making it difficult to illuminate the entire DMD surface. Make sure that you get a laser that produces a spot. There are semiconductor lasers that produce a lines and crosses and again they won't cover the entire surface of the DMD. Semiconductors come in different diameters varying from about 5mm to slightly over 10mm. You want to choose a size that matches the optics described below. Most of the major electronics suppliers carry a wide range of lasers.
While red lasers are relatively cheap, that isn't the case for green and blue lasers. If we want a full colour display we will need both a green and a blue laser. Searching the web I've found them for about $90 each. Putting two lasers into this design is fairly easy, adding the third is much harder. But, that's something for the future.
The beam coming from the laser is too small to illuminate the entire DMD so it must be magnified. There are several ways of doing this, but the most convenient way is to use a microscope objective. While an lens system could probably be constructed for less cost, this is a quick way to get something that works. For the current design we are using an Edmund Optics #43-903 (note, the part number in the paper is wrong), which provide 10x magnification. The laser we are using is a Quarton VLM-650-04-LPA-ND available from DigiKey electronics. The diameter of this laser is 6.55mm and hole at the non lens end of the microscope objective is slightly larger. This provide a nice tight fit for the laser eliminating the need to glue it to the support. The other common laser size is 10.4mm, which would require support for this objective. There are other sources of microscope objectives on the web, and the least expensive quality ones sell for about $50. A complete assembly using a 10.4mm laser is shown below. The other pictures on this page show the 6.55mm laser with the same objective.
The final thing we need to do is assemble all the components into a rigid structure that we can use. This is particularly important for the laser, since the leads can easily become detached from the body as shown below.
Ideally the display should have a custom case, possibly produced by a laser printer. This is something we are working on, but for prototyping something more flexible is required. We need something sturdy, but easy to use and shape. For our work we have used a standard Meccano set that is available at many toy stores. This is basically a set of metal parts that are held together by nuts and bolts. There are many ways of putting this together, the approach that we have used is shown below.
A Molex connector is used to connect the laser to a battery compartment. The laser runs on 3V and the battery compartment has two AAA batteries. Both the Molex connector and the battery compartment are attached to the device using Velcro. The following picture shows the bottom of the display with the battery compartment attached.
