More of this Feature • Part 2: Power, Completion     Related Resources • Keyers Reviewed • Build a WearComp • Other Products

First select a mainboard for the computer, and the various peripherals you would like, and a metal box of the right size to house it. Here I am choosing a TME-INC (made in Toronto, Canada) main board because it has onboard SCSI for my high resolution wearable camera system, which operates over the SCSI channel. This camera system provides images of resolution 3072x2048 pixels, which are processed using the VideoOrbits method, to give higher (e.g. 10,000 pixels across or so) resolution composites, so that pictures taken with the apparatus can exceed the resolution possible with photographic film (e.g. even when large format films are used).

The system has four 18 gigabyte hard drives, giving it 72 gigabytes of storage, and, with striping across multiple hard drives in an array, may provide high speed writes. Two RAM slots provide 512 megabytes of RAM, which is useful to buffer the pictures during processing.

I put together a number of different kinds of systems, some smaller, for smaller images (e.g. 640x480 image capture, etc.) and some larger (e.g. for larger image capture, processing, etc.). This particular computer system described here is not a joke, so if you're interested in packing serious heat, read on!

The entire system fits inside a Mountainsmith belly bag called a ``TOUR PACK'' (the smaller of the two main belly bag sizes offered by Mountainsmith, e.g. not the larger one that is called a ``DAY PACK'').

I chose a soft iron housing, rather than an aluminum housing, in order to have better shielding (e.g. shielding from magnetic fields as well as electric fields).

The first step is to drill the holes for the mounting of the components. I begin by marking out a piece of cardboard (drawing the outline of the mainboard on it) and making markings on the cardboard where the holes are to be ultimately drilled in the metal housing:
marking cardboard
marked cardboard

Once I have marked the cardboard, I cut out the cardboard:
cutting cardboard
Now I have a piece of cardboard that is the same size as the mainboard, and I can position the cardboard in the metal housing I have chosen, and I can now position the other components around the mainboard, and experiment with positioning of these components without risking damage to the mainboard. The cardboard therefore acts as a surrogate mainboard in the planning stages, and I can feel free to use a nail or punch to mark the metal housing, using the mainboard as a template. One must be certain to also plan for all the cabling and wiring into and out of the metal cabinet. For example, the SCSI cabling is 68 pin SCSI, very delicate ribbon cable, so it must be brought out carefully. Once all the planning stages are complete, and I have marked all the hole locations, I begin with a punch, to punch reference marks for the drill, so that the drill will later not slip out of place. I use a hammer to make the reference marks with the punch:
punch

Next I drill small marker pilots. I don't drill all the way through because I don't want to get any metal shavings on the workbench:
drill

Once I have lightly drilled the marker pilots, approximately halfway through the metal, I then take it outside into the hallway, away from the workbench, and drill all the way through:
drilling

If the weather is nice, I sometimes also do this outside, since the drill is battery powered. After all of the drilling, I file down the edges of the housing, so there are no sharp edges (most project cases, hobby cases, and various other metal boxes have sharp edges, and I prefer to file these down so that I don't cut myself on the metal as I am working on the system. Since it is an experimental system, it will often no doubt find itself being operated while opened up.) Also, I file down all the burrs and rough metal edges created by drilling the holes.

DECON

It is very important to thoroughly decon (decontaminate) the metal housing, to make sure all the metal filings are cleared away. Even just a small metal shaving, or metal dust, can short out the circuits inside the computer. Therefore, I take extra precaution with the decon. Every building complex has its best and worst faucets, and I have found that the rightmost faucet in Sandford Fleming building, room 2107, is what I believe to be the best in the building. For some reason it doesn't have a flow restrictor on it. In today's world of restricted faucets, showers, and the like, a high pressure faucet is worth making not of, if you ever find one. Preferably this should be away from the work area. I also need to be careful not to splash my other rig (the one I am wearing to documented the assembly of this rig). After this light decon, I'll peel down and do some serious decon:
decon
After serious decon, comes drying time. In the drying area, I prefer to use paper towels and scrub thoroughly, in case there are any particles of metal dust that might have been harboured by the metal housing:
dry
Small inaccessable areas of the housing can still harbour contaminants. Careful decon is therefore strongly recommended.

Next page Modifications, Power Supply and Finishing Page 2

Previous Articles