FABulous christine

Monday, October 18, 2004

tom robbins

in the center of a marble-top desk, directly under a crystal chandelier, sitting alone on a silver tray, was a large, raw beet. the beet must have been out of the groun a week or more, for it had the ashen exterior of a cancer victim. yet, when struck at a particular angle by a flicker of candlelight from the chandelier, its heart of wine-drenched velvet shone through.

anecdotes

[all things chandeliers]

allusions, inspirations, references:

lighting

some contemporary chandeliers in frankfurt; some projects with LEDs

chandelier innovation

With “Scheherazade” Schonbek developed an entirely new way to construct a beaded chandelier (patent pending). Each strand of beads is encased within four slender strands of metalwork, so that the beaded arms and scrolls appear to float.

Robert Frost

I had a vision of them put together
Not like a man, but like a chandelier.

Jane O’Reilly

I used to join the murmurings about “Where are the qualified women?” As we murmured, we would all gaze about the room, up toward the chandelier, into the corner behind the potted palm, under the napkin, hoping perhaps that qualified women would pop out like leprechauns.

Sam Moore

Radio tried everything, and it all worked. It invented a new kind of singer whose voice wasn’t even loud enough to carry across a hotel bedroom, and Americans, as it turned out, would rather hear these “crooners” than any big-bellied tenor who ever shook an opera house chandelier.

love in a cottage, willis

But give me a sly flirtation
By the light of a chandelier— 10
With music to play in the pauses,
And nobody very near;
Or a seat on a silken sofa,
With a glass of pure old wine,
And mamma too blind to discover 15
The small white hand in mine.

the story of the white blackbird, de musset

Chaffinches, linnets, and robin redbreasts had grouped themselves lightly on some pruned branches like crystals on a chandelier.

vanity fair, thackeray

O heavenly!” said Miss Sharp, and her eyes went from the carpet straight to the chandelier.

hop-frog, poe

The grand saloon in which the masquerade was to take place, was a circular room, very lofty, and receiving the light of the sun only through a single window at top. At night (the season for which the apartment was especially designed) it was illuminated principally by a large chandelier, depending by a chain from the centre of the sky-light, and lowered, or elevated, by means of a counter-balance as usual; but (in order not to look unsightly) this latter passed outside the cupola and over the roof...

At his suggestion it was that, on this occasion, the chandelier was removed. Its waxen drippings (which, in weather so warm, it was quite impossible to prevent) would have been seriously detrimental to the rich dresses of the guests, who, on account of the crowded state of the saloon, could not all be expected to keep from out its centre; that is to say, from under the chandelier...

While the tumult was at its height, and each masquerader attentive only to his own safety (for, in fact, there was much real danger from the pressure of the excited crowd), the chain by which the chandelier ordinarily hung, and which had been drawn up on its removal, might have been seen very gradually to descend, until its hooked extremity came within three feet of the floor...

Soon after this, the king and his seven friends having reeled about the hall in all directions, found themselves, at length, in its centre, and, of course, in immediate contact with the chain. While they were thus situated, the dwarf, who had followed noiselessly at their heels, inciting them to keep up the commotion, took hold of their own chain at the intersection of the two portions which crossed the circle diametrically and at right angles. Here, with the rapidity of thought, he inserted the hook from which the chandelier had been wont to depend; and, in an instant, by some unseen agency, the chandelier-chain was drawn so far upward as to take the hook out of reach, and, as an inevitable consequence, to drag the ourang-outangs together in close connection, and face to face.

fundevogel, grimm

"Fundevogel, never leave me, and I will never leave you." Fundevogel said: "Neither now; nor ever." Said Lina: "Then do you become a church, and I'll be the chandelier in it."

the little white bird, barrie

There were many oil-paintings on the walls, mostly without frames, and I must mention the chandelier, which was obviously of fabulous worth, for she had encased it in a holland bag.

I looked around me despairingly, and my eye alighted on the holland covering. "There is a fine chandelier in that holland bag," I said coaxingly.

She sniffed and was raising an untender hand, when I checked her. "Forbear, ma'am," I cried with authority, "I prefer to believe in that bag. How much to be pitied, ma'am, are those who have lost faith in everything." I think all the pretty things that the little nursery governess had made out of nothing squeezed my hand for letting the chandelier off.

the master and margarita, bulgakov

Then, he began crawling up the wall, ripping the wallpaper with his claws, and in two seconds he was high overhead, sitting on the metal curtain rod. In a flash, hands grabbed at the curtain and tore it down together
with the rod, letting the sunlight burst into the darkened room. But neither the fraudulently revived cat nor the primus fell down. The cat, still holding on to his primus, managed to swing through the air and land on the chandelier that was hanging in the center of the room."

The cat escapes with his primus stove and more havoc ensues.

proposal

[half-way mark.]

ideas on final projects:

chandelier of change

a bit fascinated with lighting; using alternative materials and mechanisms to create a chandelier with a twist. inspired as well by the evans exhibit up at the lvac. the mix of the antiqated style + digital capabilities.

inspiration: the intricate, gorgeous chandeliers at harvard's dudley house dining hall. they seem so capable of magical things.

plan: set up LEDs as the pinpoints of light, like the array of candles. put pressure sensors to detect air speed, so the user could put out whichever candles they wanted by definitively blowing them out. to light them, you could hold a lit candle to an unlit one, and the light would be transferred over as with a flame. you could also light one by itself by closing your hand [squeezing?] over it and then uncovering the light. the hand holds a bit of mystery.

the intensity of the light would be manipulated by either blowing softly [a progressive dimming] or cupping your hand around it as to protect the flame [a progressive brightening].

other than the manual lighting interface, the lights could throb in time with a user's heartrate - connection via pulse reader. could also assign each light to a specific star, and twinkle them with varying intensity on the light with real-time cosmological data. perhaps i could hook this up with an email correspondance database, with the lights forming an ambient device that notifies which people you've had good response rate with, and which rapports are dying away [light fading].

want to infuse a bit of unpredictability, surprise.

materials: programmed microcontroller, sensors, LEDs (3-way, hopefully), data input devices, construction of acrylic, aluminum, and plastic

the dancer

i like the idea of using mixed media, a la degas with his dancer. the 3d printers take forever and you have a tradeoff between color and strength. however, could do more with creating more organic pieces by printing in ABS and fusing bits of fabric, texture, fur, found objects, onto and around it.

three thousand dimensions

[foray into the monstrous]

happily [or not so, for some], the stratasys printer got back online by the end of last week. john gave us a little primer, which goes a little something like this:

stratasys printing:

takes input as STL files (apparently it likes ASCII mode, whereas the zcorp has a predilection for binary files). unless you want your thing to print for a day and a half (quite literally), keep the z-values low, and your thing small. in the insight software, scale your object using STL -> scale. on screen, navigate through menus and things using the << >> buttons, and make sure settings are what they should be (as posted). click the green flag icon to start conversion of the STL file for the printer. click the button with the little horizontal lines to sift through the stack of layers that are created from the process to see if everything looks good with the binding material and if there's wiggle room where you intended it to be. to toggle between the object and support, go to STL -> display.

start communicating with the printer! click thing on screen that looks like a box talking (ok, ok, it's an arrow) to another bigger box.

now, to the machine: push ENTER and make sure that the support material temperature reaches 235. ensure the foam pad is agreeably smooth and secured by pins. now, the purging process to make sure the heads are smooth as silk. push LOAD when you see MODELING material appear on the LCD. the head should jog forward and the nozzle starts going. when the thread that emerges is whitish and smooth, push LOAD again to stop. to clean the support nozzle, push MODEL SUPPORT, and then do the same thing with the LOAD buttons. when done with this whole thing, make sure the modeling nozzle is back in the lower position.

ah, how cathartic. everyone needs a good purging.

pushing PAUSE puts it to absolute (0,0) in the (x,y) direction. you can press arrows to reset it if necessary. to adjust the z, press Z-AXIS and then use the up-down arrows. you want to bury the modeling tip about 1mm in the foam (ever-so-slightly below the surface).

make sure the printing is ok: the support material should be piped out in an outline first, NOT a raster. if this is not the case, make sure all is well; press PAUSE. on a good day, the support draws the outline and then fills in by rastering.

the nozzle heads we use are the ones with 2 rings on them, a T12 model. claim to have 1mil resolution; seems like for moving parts, the tolerance hovers around 20mil.

we had initial problems of the stratasys suddenly shutting off repeatedly. cause was ascertained to be overheating of the interior envelope, so leaving the door ajar seemed to do the trick.

the experiments

so although the zcorp dazzled with its petite stature and color capabilities, i didn't want to deal with particles in the lung and shattering [hearts] pieces. i'll play around with it soon enough, but set my eyes on the stratasys because of its durable ABS. or so i first thought.

spent some time playing around with 3d studio max and fell in love with it's intuitive-like-blender-is-not quality. played around with the primitives and thought an abacus might be something cool that dealt with interlocked moving parts and become something remotely useful. also made a bracelet to match (resembles those tasty candy necklaces we all know and love) for fun, and tried out a torus knot to see how it'd print.

design files

time is of the essence. on the stratatys, two objects:
my abacus (resized to 50%) 4" x 2" x 1/4" (relatively flat) and the knot about 2" x 2" x 2". they're laid side by side. the estimated printing time is 13 hrs. (!) the reason it took so long was the height of the knot.

i wasn't there to witness it, but i heard my abacus was so fragile it snapped in pieces while support material was removed from it before the bubbling bath. the remains:



reprinted the abacus at a larger size. much better; the beads resemble cheerios. however, the plastic is quite noticably stringy around the beads. the rectangular frame, however, is sturdy and clean-cut. my hunch is that the printer doesnt do terribly well with small areas with high curvatures. the torii of the beads are probably curvier than the printer can handle to print cleanly.

and the knot? it turned out gorgeously. check it out!

Tuesday, October 05, 2004

dreams of white

[learning how to use the 3D laser scanner and printer]

for next week: take anything 2D or 3D, provide input, and print output. [the possibilities loom.]

here are the best attempts to gather the knowledge i've absorbed in the last two days regarding the 3D machines. so fancy; this is the moment i've been waiting for.

the dreaded blender resurfaces for texture mapping bitmaps onto mesh curves. the plan: learn 3d studio max this week and see if it's easier to work with. but in case i bug out...

blender [newest version]:
- create a mesh [e.g. uv sphere - 12 segments, 12 rings].
- uv face select mode
- uv image editor mode
- [on right screen] image : open image
- U_key: sphere
- texture
- A_key: shift-right-click the polygons
and then, exporting as VRML 2.0 preserves the color mapping.

minolta laser scanner:
- SCSI connection, so turn scanner on, then turn computer on
- geomagic studio software
- i found this software tutorial and this other tutorial online
- show image window -> show streaming video
- can either autofocus or input manual distance and power
- each scanning pass obtains 300,000 points of data [wow!]
- to clean up, points -> reduce noise, points -> wrap, polygon -> fill holes
- to smooth edges, tools -> make boundary edges -> path fit
- to simplify, polygon -> decimate
- to stitch together several scans, callibrate z-axis, tools -> registration
save as VMRL 2.0 file!

so there are two printers at our disposal: the stratasys, which takes in STL files and prototypes in heavy-duty monochrome ABS plastic, and the zcorp, which takes in STL or VRML files and models in color using a special plaster. sadly, at the moment, the stratasys is out of commission. the zcorp prints to 0.004" resolution.

zcorp printing:

[to fill the trays, can either do automatically or manually.]

auto: go to online mode, and do 406service -> fill bed -> start. the printer goes and does its own thing for a couple minutes, spreading the plaster smoothly from the feed bin to the build bin. use the vacuum on the side to suck up the excess dust. to get the particles underneath the head, go offline, press spread to move the head, and suck away.

manual: make sure you're in offline mode. even out contents of feed bin with special flattening tool. move the build bin to about quarter-inch below the top lip. spread. raise feed bin to about quarter-inch above the top. spread. raise feed bin. spread. keep spreading until things look smooth. lower build tray to about quarter-inch below the lip. vacuum excess. [keep things pretty and neat while working; the white dust gets everywhere.]

go to 406service -> unpark and use deionized water to wet the wipers and the cups. dry only the cups with a paper towel.

on with the software! zprint!

when using the zprint software for the zcorp printer, should always import your files. the most efficient way to use the space within the printing envelope is to place the objects on the bottom; keeping z-values low is key to saving hours of plaster laying. good idea to buddy up with friends to lay jobs simultaneously on the printing bed.
- transform -> justify to line up to area envelope
- do print estimate (quicker for monochrome)
- problems? try edit->invert normals
- default with spatial interior
- want a shell? take boolean subtraction of smaller shape from larger one

click the button for 3D print setup and choose either setting of ZP102 [drier] or ZP56-102 [wetter, better color]. if you want monochrome printing, choose 406service -> change binder supply. the best part of this is an email notification service that tells you when your thing has finished after 31.7 hours! edit -> preferences to enter email. :)

post-printing, take dusty object over to the vacuum station and scrape, wipe, suck, and excavate to heart's delight. post-post-printing, put into the oven machine to heat and wax, which strengthens and preserves the plaster model.

the coolest models to come out of the 3D printers are those with integrated free-moving parts, such as a ball-bearing system, a time-watch mechanism, or a ball-and-socket connected vertebrae. more skeletal-inspired ideas: fixed joint, hinge joint, pivot joint, sliding joint. i wonder if i could make a rough metronome. found page of cool historical instruments.

Monday, October 04, 2004

microcontrolling the world

[milling via modela, surface-mount soldering, tied up in knots, and windows assembly]







my procedure for getting this going on windows follows. i was a little hesitant to install linux from scratch on my computer, because that might invite more trouble, especially in the limited time window i had for the assignment.

download avra with the include files tn15def.inc and tn26def.inc, download uisp [on the download page, you want the link that's
like the 3rd down [ uisp-20040311.tar.bz2 ; i used winrar to extract the
files], download python, download pyserial, pyparallel, giveio, and pywin32 (all on sourceforge)

you need cygwin to compile uisp; when installing cygwin, make sure you install the auto-make and gcc components [in that dialog box during setup, click 'view' in the upper right and the components are in alphabetical
order]

then go in cygwin, get in the uisp directory, and type ./configure. after
that, type make. and then make install. things should go ok.

you have to change the paths in windows; to to my computer, right-click,
properties, advanced tab, and then environmental properties. edit the path
settings (on the bottom) and add the directories that python, avra, and uisp
are located (separated by semi colons)

alright, now you have to check what your parallel port and serial port are
called. go to my computer - properties - device manager. check out the
ports. my serial port was COM5 [although every time my computer started up it would rename the port to COM4 or COM6, you want to check each time]; my parallel port was 0378 (click on lpt1 and do properties, resources tab, the number should be on top; all of ours were 0378 so i think that's the default)

edit your rx.py file so that the serial port reflects your COM port. mine
has '/COM5' instead of '/dev...' whatever is there.

now... run->cmd.exe. go to your directory with the rx.py file. type python
rx.py. keep this window running in the background. [this is where the hello
world stuff should display when it's running]

open another cmd window. go to directory with your hello.asm file. type avra
hello.asm -- this compiles the code and creates a hello.hex file.

type the uisp command that neil has to load code, except change the parallel port to reflect yours; mine looks like:
uisp -dlpt=0x378 -dprog=dapa -dvoltage=3 -dt_sck=50 --erase --upload if=hello.hex

once you type that, it should say something like 'uploading; flash
installed' or something to that effect, and you should see random letters
streaming down on the rx.py window (or hello world if you're really lucky).
most likely you'll get some weird error. check to make sure all the cables
are really connected, and shove those serial cable wires even tighter.
anyway, it should work if the connections are good.

and then, if you find you're getting gibberish letters, edit the hello.asm
file and change the b value. every time you edit it, save, and then
recompile using avra. load onto chip using uisp and keep trying til it
works.

and voila! here's the beautiful evidence...



metal crimp madness!

it took me about an hour to do my first one; then realized you're supposed to thread the wire through, not shove it in from the top! :)

i just heeded raffi's advice; strip a wire, twist the strands, then thread
thru the crimp. make sure there's rubber wire covering going through the
first (the largest) pair of clips. take needlenose pliers or something and
just squeeze the clips to tighten around the rubber (like tongs). put a dot
of solder in the middle of the crimp to secure the wire inside (although i
think mine are not really that secure). i was careful not to put too much
solder because i know that's diastrous too. cut off the excess wire from
the end.

then, bend off the crimp from the metal tape and stick into the serial
connector. i used pliers to shove the thing in completely. make sure the
widest part of the base is in.

i think the best thing is to check every time that all the crimps are in the connector before trying to upload/download to the chip. they arent the most secure things; i thought the metal crimps would catch onto something inside the plastic serial connector to help them stay in but there's always danger of them slipping ever-so-slightly out.

the divisions of time and space

[lasercutting, ho!]

will fill in the details later. but for now, a picture:




there's a 40W bulb inside; the evanescent glow is quite nice.