by Richard F. Drushel (drushel@apk.net)
One of the best things I ever built for my dedicated ADAM system was a wooden stand for the monitor. While it's possible to set a monitor directly on top of the ADAM CPU (if it isn't too heavy), it's really not a good idea--the CPU case isn't so strong, you can get electrical interference between the ADAM video circuitry and the monitor, and it's really inconvenient if you need to open up the CPU to get at expansion cards.
My monitor stand design is basically a 4-legged table which fits over the CPU, with about 6 inches of space between the top of the CPU and the wooden table top (thus allowing access to the internal expansion slots). The space between the legs at the sides allows access to the power cable and disk drive ADAMnet cable (left side) and expansion port devices like serial modems (right side). To provide for convenient electrical hookups, I also have a 6-port powerstrip attached to the back, with a 6-foot cord. This way, wherever I take my ADAM, I only need to find one 120-volt outlet and I'm ready to go.
As I mentioned a few weeks ago, I've been slowly reorganizing my basement so that I can have as many working ADAM systems set up as possible, for use by my kids and for our users group, B.A.S.I.C. The first ADAM system I set up was not my "experimental" system, but I gave it the latter's wooden monitor stand. In order to set up more ADAMs, I would need more monitor stands; thus, I'd have to build them. This week, I built another one.
Here are abbreviated, no-picture plans for my monitor stand. The text should be pretty self-explanatory, but if anybody wants something fancy like a measured drawing, please let me know. Note: all dimensions are in feet and inches; you progressive Canadian metric types can convert :-)
One 8-foot 2x4 (make sure it's straight, no twists or bad knotholes), one 4-foot length of 0.75-inch square hardwood moulding, four 6x8 inch metal angle brackets (the kind that are stamped from steel and have a U-shaped channel in them, not just a right-angle bend of bar stock), 36 1-inch x 8 steel wood screws. You can get fancy and stain/varnish or paint the wood if you like, but bare wood travels better when piled into a van full of other computer stuff on its way to an ADAMcon :-) Optional: a 6-port powerstrip with or without AC and/or phone line surge protectors, circuit breaker, etc.
Hand or electric drill, bit to match wood screw diameter, saw (you can use a carpenter's saw or backsaw if you're really good, miter box with backsaw is better, best is radial arm saw or other console-mounted saw), large screwdriver (slotted or Phillips, to match head of wood screws), steel measuring tape or large carpenter's square, 150-grit sandpaper and sanding block or electric sander.
Two 26-inch beams supported by two 10-inch legs each. Joint of beams and legs reinforced with angle brackets. Two assemblies connected by four crossbeams underneath main beams at an (outside) width of 12 inches. See the ASCII artwork below.
Measure the lengths of the two beams (26 inches each) and the four legs (10 inches each) on the 2x4. Remember to leave the width of the saw blade (kerf) between the measurements, or some of the pieces will end up too short. Cut the pieces. Sand them down smooth. If you cut them by hand, make sure that the edges are square (stand up the pieces on a flat surface and see if they veer to one side or another); sand them flat if not. (This can be a major chore if you don't have a power sander.)
Lay out the two 26-inch beams side-by side, good wide side down. Stand up the legs onto the beams, one leg at each end. After you're sure that everything is lined up and square, position the angle brackets at each corner, 8-inch side down, 6-inch side up, centered on the beams/legs. You can get fancy and draw centerlines etc. if you want to, but by eye is good enough. With a sharp pencil, trace the screw holes onto the wood, being careful not to disturb the alignments. This can be tricky if you have shaky hands. My angle brackets have 6 holes per bracket; your mileage may vary. Use a hammer and a nail/centerpunch to mark the center of the holes, then drill them out. The drill bit has to be a little smaller than the shaft of the wood screws you're using, but not so small that it's too hard to put in the screws. (Don't make them too big, though, or the screws won't bite into the wood and they won't hold.) Don't drill all the way through, either! About 1 inch deep is right. Now screw everything together. If you are of average strength but are having a hard time putting in the screws, your pilot holes are too small; try making them a little bigger with some up and down strokes with the electric drill. Once they're assembled, turn the two assembiles over and make sure that they are level and aligned. You might even try a test with your intended monitor to make sure that it doesn't bend under load.
Cut two 12-inch pieces of the 0.75-inch square moulding (again, remember to allow a kerf thickness between the pieces when measuring). Lay the completed leg/beam assemblies top down (legs in the air) side by side, 12 inches apart. Use a carpenter's square, a wall, etc. to keep the ends (legs) parallel. Lay the two crossbeams across the main beams, each one just past the tip of the angle bracket. Mark two holes at each end of each crossbeam, about 2 inches apart and about 1 inch in from each edge of the 26-inch beam. Again, you can measure exactly or lay it out by eye. Take the crossbeams aside and drill out the holes (you don't want to disturb the setup on the floor). After they're drilled, bring them back over, reposition them, and mark the holes onto the beams with a nail/ centerpunch. Drill out these holes. Finally, attach the two crossbeams with screws. Turn the almost-completed stand over and make sure it's still square and level.
Turn the stand top down, legs up again. Measure the distance between the inside edges of the angle brackets, across the center gap, back in the corner where the legs meet the beams. Depending upon how accurately you positioned the angle brackets, these lengths may not be the same at each end, but it's not too critical. Cut two more pieces of the 0.75-inch square moulding to match what you just measured. (Mine ended up 7 inches long apiece.) Position one in each corner across the center gap. Mark one hole at each end of each crossbeam; drill the holes. Now comes a tricky part: depending upon how big your drill is, you may or may not be able to get it in position to drill straight down through the existing holes in the crossbeams. If this is the case, you have no recourse but a heavy-duty screwdriver and elbow grease to get the screws in that last quarter- inch or so. In the best of all possible worlds, you could lay out all the screw holes before anything was assembled, drill them out, and then put it together and have everything line up correctly. In my experience, unless you have $20K of shop tools, you can't get it right by predrilling. You could unscrew the legs, but for me it was easier to just screw in the crossbeams. Screw them down, turn the stand over, and admire your handiwork!
Every powerstrip I've seen has some kind of holes for mounting it by screws. There may be a set of flanges on the outside of the case, but usually there are miserable lightbulb-shaped holes in the back: screw heads fit through the round part of the hole, then you slide the powerstrip right/left/up/down so that the shafts of the screws go into the slot part of the hole. In the latter case, the hard part is getting screws mounted so that they have the right spacing. Don't even try to measure them; instead, make a pencil rubbing of the back of the powerstrip onto a piece of paper. Lay a piece of paper over the back, and scribble lightly with the side of a soft pencil all around the holes, being careful not to move the paper (tape it down if you like). When you're done, you have a mirror image of the slots. Mark the centers of the round parts of the holes, poke through the centers with the pencil point, then turn the paper over and use it as a template for laying out the holes on the back of the monitor stand. Put in the mounting screws (predrilling isn't usually necessary, but you can if you want), and screw them in until there's just enough gap between the heads and the wood to equal the thickness of the metal/plastic back of the powerstrip. Make a test fitting with the screws a little loose; when you're sure it will fit, tighten the screws down a little at a time and keep refitting (press on, slide) until it's snug.
Congratulations, you're done!
Here's some ASCII art (not to scale) to show the general arrangement of the parts, in case my verbal description wasn't clear to you:
+----------------------------------------+
| | <2"
| |
+---+--------+-+-----------+-+-------+---+
| |@@@@@@@@+-+ ^ +-+@@@@@@@| |
| |@@@ ^ ^ | @@@| |
| |@@ | | | @@| | FRONT/REAR VIEW
| |@ | | | @| |
| |@ | | | @| |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | \ | |
+---+ | \ \ +---+
^ | \ \ <2"
| | \ \
| | \ \
10" angle 12" 26"
leg bracket cross beam
beam
+---+--------+-+-----------+-+-------+---+
| | | | | | | |
| |@@@@@@@@|.| |.|@@@@@@@| |
| |@@@@@@@@|.| |.|@@@@@@@| | <4"
| |-+ | | | | +-| |
+---+.|------| |-----------| |-----|.+---+
| | | | | | | |
| | | | | | | |
| | | | | | | |<-------\ BOTTOM VIEW
| | | | | | | | \ (LEGS POINTING UP)
| | | | | | | | \
+---+.|------| |-----------| |-----+.+---+ \
| |-+ | | | | +-| | \
| |@@@@@@@@|.| |.|@@@@@@@| | |
| |@@@@@@@@|.| |.|@@@@@@@| | <4" |
| | ^ | | | | | | |
+---+----|---+-+-----------+-+-------+---+ |
^ | ^ ^ <2" |
| | | | |
| | \ \ |
| | \ \ |
10" angle 12" 26" . is a screw approx. 7"
leg bracket cross beam crossbeam
beam
+-----+ +-----+
| |<2" | |<----\
| | | | \
+-----+-------+-----+ \
| |_______| | \
| | ^ | | \
| | | | | | END VIEW
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | Note: "2x4"s aren't really
+-----+ | +-----+ | 2 inches by 4 inches any more;
^ | <4" | they're more like 1-9/16 by 3-7/16
| | |
| | | :-(
10" end 26"
leg crossbeam beam
Dale Wick (dalew@truespectra.com) asked me to send him pictures of the eight ADAMcon banners which I photographed at the ADAMcon VIII banquet. (Traditionally, all the banners are hung up in the banquet room during the banquet; I hope that Bob Slopsema (72117.3003@compuserve.com) can find enough wall space for them for ADAMcon 09!) Dale wanted to put the pictures up on the Coleco ADAM Home Page (whose URL I'd give you except that the Coleco ADAM Home Page has moved and I don't know where), which means that somehow the photographs would have to be converted into graphic image files. Since I have access to the necessary equipment at my lab, I agreed to convert the photographs to image files.
The basic strategy is simple: use a device called a flatbed scanner to scan the photographs and produce a high-resolution color image file, optionally do some image processing on the image, then convert it into a file format compatible with common Web browsers like Mosaic or Netscape.
The typical modern scanner is capable of scanning images at 600 dots per inch (dpi) spatial resolution, with each dot (pixel) represented in 24-bit "true color" (i.e., 8 bits each for red, green, and blue light). It can also produce images in 8-bit color (3 bits for red and green, 2 bits for blue), 8-bit greyscale (0=black, 255=white, shades of grey in between), and 1-bit line art (a 0 bit=white, a 1-bit=black).
The greater the spatial resolution and more accurate the color, the bigger the resulting image file. For example, a 1-inch square, 8-bit greyscale image at 72 dpi is 72 x 72 = 5,184 bytes. The same 1-inch square image at 300 dpi and 24-bit color is 300 x 300 x 3 = 270,000 bytes! The size at 600 dpi/24-bit color is left as an exercise to the reader :-)
The larger the image file, however, the longer it will take to download to a computer for viewing. Assuming a 14400 bit-per-second (bps) modem at 8 data bits, no parity, and 1 stop bit, the 72 dpi greyscale image mentioned above will take (9 x 5184 / 14400) = 3.24 seconds to download, while the 300 dpi "true color" image will take (9 x 270000 / 14400) = 178.75 seconds!
The solution to minimizing image transfer time (aside from faster and faster modems) lies in the technique of image compression, i.e., the image data are encoded in a format which takes up less space than the original image. Some methods of compression do not alter the original data, such as the Lempl-Ziv-Welch (LZW) algorithm used in the CP/M CRUNCH.COM and earlier versions of the MS-DOS PKZIP.EXE; the compression algorithm is completely reversible. Other methods of compression, such as that used in JPEG images and MPEG digital movies, achieve great size reduction ratios by changing the original image through spatial averaging of colors. These latter forms of image compression are termed "lossy" because some of the original image details are lost during the compression. In the case of JPEGs and MPEGs, however, it is possible to specify different tradeoffs between image size and image resolution; you can pick how lossy you want the final compressed image to be. A final method of image compression is called palettization or indexed color: instead of having an infinite number of colors available, you pick a small subset (the palette), give each palette color an arbitrary number (the index), and map every pixel in the image to the index of the color in the palette which is closest to the original color. Typical GIF87A color images use this method of size reduction: part of the file specifies a table of 256 different colors (in 8-bit or 24-bit format), and the image proper is just an array of indices 0-255; the program reading the GIF file makes the appropriate color substitutions from the palette as it draws the image on the screen. An interesting sidelight about GIFs: not only do they use palette color, but their image array fields are themselves compressed with the LZW algorithm, thereby achieving an even greater size reduction.
Last week I finally finished out the roll of film that had my ADAMcon VIII pictures on it and got it developed, so at last I had the photographs. This last Saturday night, after the kids were put to bed, I went into the lab to make image files. Here's how I did it:
I scanned the original photographs at 300 dpi 24-bit color and saved them as LZW-compressed TIFF files. TIFF was one of the first image file formats developed and is good for lossless images. The scanner software (Ofoto for Apple Power Macintosh) has some nifty features like prescan (to see quickly what you've got, to make sure it's lined up, etc.) and image cropping (you can specify an arbitrary rectangular subset of the prescanned image, so you can just grab the parts you're interested in). These images were pretty huge, most 3-6 megabytes in size, so they definitely would have to be compressed in some way for practical viewing and downloading on-line.
I opened each TIFF image in Adobe PhotoShop 2.5.1 for Macintosh. This powerful image-enhancement program allows you to do all kinds of photographic retouching, special effects, you name it--including color-correction. Often you will get some color differences in scanned images compared to the originals. The image might be too green, or not have enough red, etc. This is a variable for each individual scanner. Also, you might be scanning an old photograph whose colors have changed with age, and you'd like to restore the original colors, contrast, etc. PhotoShop has sophisticated tools to do color and contrast corrections. The scanner I used made the images come out darker than the original photographs and also a bit too red. I toned down the red and turned up the blue to compensate.
Next, I converted the images from 300 dpi to 72 dpi without changing the actual image dimensions. This is a lossy compression which reduced the TIFF files to about 150K each. (If I had gone from 300 dpi to 72 dpi without constraining the image dimensions, the images would just have been displayed larger on the screen, with no shrinkage of the data.) This size is tolerable for a single image or so on a web page, but if you want people to look at all the pictures, 150K is too big.
The final compression step was to convert the 72 dpi 24-bit color TIFFs to 24-bit color JPEGs. I selected 75% image quality, 25% image size as my lossage parameters (higher means more of that parameter), and the final JPEGs were about 15-20K. That means only 10-15 seconds of download time each, and that's acceptable.
This morning I put all the final JPEGs up for download and sent mail to Dale Wick that he could grab them at his leisure. So can you! All the files are in the /ac08 directory and have the extension .jpg. The files named ac01 through ac08 are the eight ADAMcon convention banners (basic is the B.A.S.I.C. user group banner). The files named banq01 through banq05 are banquet table shots. The names of the rest are self-explanatory, except for rfd-jcd: that's a special treat picture of me and my wife Joan (am335@po.cwru.edu) from 1995. Let me know what you think of it :-)
I case you're wondering why I made JPEGs instead of GIFs, I tried making some GIFs and really didn't like the way they came out; they were really spotty and over-palettized, especially the banner pictures.
[entering soapbox mode]
Just a final philosophical note: I have never felt bound by the stricture that "A *true* ADAMite only uses an ADAM; he never uses an IBM or an Apple or anything else; if he can't do it with the ADAM, he shouldn't be doing it." I used to hear this frequently in the early 1990s. Throughout my ADAM career (dating from 1988), I have felt free to use whatever computing resources I felt could best help me get my work done. I have no qualms about using other platforms to get work done for my ADAM (editors, word processors, assemblers, disassemblers, etc.). In fact, if I had been limited to SmartWriter's editing capabilities, I would never have been able to disassemble EOS-5 and SmartBASIC 1.0, and I wouldn't know any of the things that I know now. If the scanner's hooked up to a Mac, use the Mac, don't lament that there isn't SmartScan 1.0 for ADAM and never be able to show others my pictures. Heck, the prospects for a SmartMosaic web browser aren't very bright, so you'll *have* to use some other computer to view the images.
[exiting soapbox mode]
See you again next week!
*Rich*
Next Article
Previous Article
TWWMCA Archive Main Page