This project is yet another node in my home’s sensor network.  It needed two main functional parts: (1) a motion sensor across the entryway to my bedroom, and (2) a switch to control the lights on my bed (my canopy bed has lots of little LEDs hanging on it).  It would communicate wirelessly with the rest of my sensor network and also allow my computer to control the bed lights.  Since it would be fairly visible, I wanted it to look pretty.  My bedroom kind of has a fairy-tale theme to it.  With a starry fairy-tale princess bed, a mushroom table, a hat vine, and dragonflies on the wall behind my bed, I decided that this sensor node should evoke the feel of a forest pond or stream.  A while ago, one of my art teachers gave me a bunch of slate because she was curious what I would do with it (thanks Dr. Ushenko!), and I decided it would be perfect for this piece.

I started with the insides.  The electronics are a lot like my other sensor nodes, plus a couple darlington transistors to control the power to my bed lights.  I designed the circuit boards and then had them made through BatchPCB, like usual.  If anyone is curious, I can go into more detail about the circuit design.

I didn’t want a plain old boring switch for my bed lights, so I disguised the switch as a cat-tail reed.  The stem is copper rod.  In the guts of the piece, this rod is used to actuate a momentary-contact switch, shown to the left of the circuit board in the following picture:

Then it was time to tackle the slate pieces.  I love working with new materials, but I was still scared that I would screw things up.  I let the project sit for over a month and finally forced myself back to it.  It turned out to be fairly easy to cut with a hacksaw.  Cutting through all that grit totally destroyed the hacksaw blade, but whatevs, hacksaw blades are cheap.  By the way, always wear a mask while cutting this stuff — silicosis is nasty, you don’t want to do that to your lungs.  I decided to carve the slate and paint it, using watercolors where I wanted washes of colors that didn’t hide the slate’s beautiful texture and ink where I wanted dark black silhouettes.  I carved it with some gouges from my set of x-acto knives.  Again, the slate dulled the blades, but those blades are easy to replace.  Actually, I found I could use the slate to sharpen them again, if I dragged them backwards through the grooves I had made; I’ll have to see if they dull quickly in the future, though.  The next pictures show a couple in-progress shots as I worked with the slate.

The top of the box containing the electronics forms the surface of the pond or stream.  It’s painted with blue and green washes and then coated with varnish.  I added a couple more cat-tail reeds, but these are just static; they don’t act as switches or anything.  I also broke up some of the slate, filed the edges smooth, and glued them to the top as rocks in the water.

Once it was all assembled, it looked like the pictures below!

My home is not all that large – it’s just a one bedroom condo.  So I don’t have room for a workshop.  But as you can tell from this blog, I like to make stuff.  Finding space for my makery’s ever expanding collection of tools can be problematic.  I recently bought a drill-press.  While trying to figure out where to put it while it’s not in use, I decided to make a mobile table for it and for my vise.  Then I can just roll it into a corner of my bedroom when I’m not using it.

I wanted the table to be stable while using the drill-press, but I also wanted it to be mobile. These design goals conflict somewhat.  So I borrowed an idea from Norm Abram (who in turn borrowed the idea from a scenic shop (for the film industry?  for theatre?  not sure)).  He describes a workbench with retractable wheels.  When the wheels are up, the table rests on its own legs and thus is quite stable; when the wheels are down, the legs are lifted off the floor so the whole thing can roll. I flipped his design around a bit.  Hinged blocks drop down to hold the wheels in place when the wheels are down.  Putting the wheels down requires kicking the blocks into place, so I have the wheel assemblies on the outside ends of the table to make kicking things into place easier.  This also made the wheelbase a bit longer, which helps a lot since the table is not all that big.  The ropes tied to the blocks just make it easier to pull the blocks up when I’m raising the wheels. Also, it’s hard to tell from the photos, but the legs are lifted about an inch when the wheels are down.

Wheels Down, front Note the hinged blocks on the sides that prevent the wheel board from flipping up.	Wheels Up, front
Wheels Down, rear	Wheels Up, rear

I also added a drawer. Roller catches keep the drawer from opening while I’m moving the table. The bottom shelf has a lip around it to keep items from sliding off when moving the table; and I mounted some pegboard on the back, so I can store even more in this table.  The whole thing is quite rigid and stable when it is on the fixed legs.  I do want to get some kind of dust cover that I can drape over the top, so it doesn’t get all dusty when it’s sitting in my bedroom.  But yay!  My drill-press and vise just became a lot more usable without cluttering my living space!

After taking a look around, I decided to locate it in my bedroom, using the space between and over a couple mirrors.  With the odd shape of the space and the presence of a faerie-tale bed and mushroom table in my bedroom, the place seemed very much in need of a vine – a hat vine!

I love the look of metal and a curvy, flowing vine just cried out for metal.  I decided to use 1/2″ x 1/8″ rectangular copper bar for the main vine stems.  After drawing a template on a very large sheet of paper, I bent the bar free-hand (with the help of a vice) since I don’t have any fancy metal scroll-making tools.  The different sections of the main vine stems are bolted together with small screws and nuts.

The leaves are wooden hearts from a craft store that I painted in various shades of green.  I used 1/8″ diameter round copper wire for the leaf stems.  The leaves come in pairs, with the leaf stems mounted through holes in the main vine stem.  I made the holes just barely big enough, so the fit is fairly tight and then I ended up using super-glue to afix the leaf stems to the main vine stems.  I’ll probably regret this (I’m not a huge fan of super-glue), but it actually seems to work on copper really well, and the leaf stems fit quite tightly before I even applied the glue.  I suppose if the bonds do break over time, I can always do something more extreme like soldering or brazing.

The leaf stems are pushed about an inch into each of the leaves.  I felt like a high-volume discount heart-surgeon as I worked through a large stack of wooden hearts, clamping each one to the “operating table” (my work space), and drilling a hole into it.  A heart looks so sad clamped to a table. 

The largest pegs the craft store had weren’t quite long enough, so I mounted each one on a fat dowel, and screwed and glued that to a square base.  The bases gave me room for the wall anchors I used to attach them to the wall and they also space the copper vines away from my walls, so the metal won’t leave any marks on the walls.  The vines hang from or rest on the pegs.  I also drove tiny nails into the pegs and bent them over the vines to better hold them in place.

And that’s it!  I now have a place to hang my hats and my bedroom now looks even more like a faerie-tale!

I finally got around to making a light-switch for the chandelier.  Weirdly, this wasn’t a huge priority – since the chandelier can be controlled from my main computer, I didn’t need to make a dedicated switch to control it.  Having a light-switch on the wall nearby, though, is convenient.  I didn’t want to always have to run over to my computer just to turn the light on.

It’s nothing all that fancy.  I cut the green acrylic for the faceplate from a piece leftover from the chandelier.  The yellow acrylic flower was an extra from the flowers I laser cut for the chandelier.  I put some silvery mirrored paper behind each of the acrylic pieces.

It communicates wirelessly with the chandelier.  Since the chandelier can be controlled by two different devices (my main computer and the light-switch), the last position of the light-switch’s dial doesn’t necessarily convey the chandelier’s brightness; so I put a ring of LEDs around the dial as an indicator.  For the same reason, I used a rotary-encoder instead of a potentiometer for the dial.  I didn’t bother getting printed-circuit boards made just for this project.  Instead I was able to hack one of the circuit-boards I had made for my general purpose sensor nodes.  It isn’t pretty since it has wires soldered directly some of the microcontroller board’s pins, but it works just fine.

Sooo, until I think of more things to change, this brings the chandelier project to a close.  I feel kind of empty now … kinda like I need a new project … maybe a weather display or some automation for my garden or a lightning detector or … 

The chandelier has been installed!

My dad and I mounted an aluminum beam to a concrete beam near my ceiling, and the chandelier hangs from that.  Once the chandelier was hung, I added the butterflies that hang from the bottom of the chandelier.  I still need to make a light switch so that I can control the light, but in the meantime, since it communicates wirelessly with the digital assistant running on my main computer, I have full control over the light from my computer.  (See Flower Chandelier, part 1.)

So anyway … Woo-hoo!  The chandelier project is almost done!  I’m very excited.  This has been great fun, and I enjoyed dragging my Dad into the project. (I suspect he enjoyed it too!  ).  I got to work with materials that I had never used before, which I always find intriguing.  I advanced my knowledge of electronics and encountered lots of interesting design issues.  And I finally set myself up with some decent CAD software, for several reasons:  (1) I needed to communicate my designs to my Dad for the woodworking parts he made;  (2) I had to make sure everything would actually fit together like I wanted;  (3) and I needed to create files to direct the laser cutter for the acrylic.  In fact I ended up doing so much design up-front, that the end result looked pretty much exactly like I had envisioned!  Radical.  It was like, in my brain, and then somehow popped out into the real world!  woah

Aaaaaaaaaa!!! Hexapus!!!!

Queue the trumpet fan-faire!  All the parts for the chandelier have been finished!  This article is a quick photo journey through the assembly of the chandelier.  Most of the difficulty in this project was designing and making all the parts.  The actual assembly was fairly easy.  All the parts are stacked onto a central rod.  The rod is threaded over its entire length and I used nuts and flanges between various parts to distribute the weight, so it’s not all sitting on the bottom-most part.

The white circular thing above the electronics compartment is an RGB LED strip that will provide accent-lighting for the parts above.

Top pieces

Finished assembly.
The main lights are on at their lowest power so that they don't overwhelm my camera's light-metering. The accent lights were set to emit a yellow light that mirrored the yellow of the acrylic flower petals.

There is a concrete beam crossing my ceiling that's near the place where the chandlier will hang, but unfortunately, that beam isn't quite where I want it, so I'll attach an aluminum beam to it and the chandelier will hang from one end of that beam.

The threads seen sticking out the top are the other side of the hook from which the chandelier's chain will hang.

Each of the six arms of the chandelier will end in an acrylic flower.  Inside each flower will be a cluster of three high-powered LEDs.  (I’m using neutral white Luxeon® Rebel LEDs in LuxDrive’s Endor Star™modules.)  These do get hot so they need heat-sinks.  The picture on the left below shows how the parts of the light assembly fit together.  The circular thing with the fins is the heat-sink, of course.  The acrylic flowers will be attached to the metal disk.  The metal cylinder with the threaded pipe-nipple screws into the heat-sink.  The white flat thing on the green wire is the LED module.  The little orange-ish thing nearby is a thermistor that I’ll be using to measure the temperature of the LEDs.  The little clear plastic piece is a wide-angle lens.  Finally, the rightmost cylinder is a hollow can that fits over the LEDs and the other metal cylinder.  My Dad drilled and tapped everything so all the parts fit together quite nicely.  The right picture below shows how all those parts once they are assembled.  Between any parts that should conduct heat, I used thermal grease.  The thermistor is attached to back of the LEDs with thermal adhesive.

Since my last article (see part 1), I made CAD drawings of all the acrylic parts and cut them out on a laser cutter.  My Dad drilled and tapped those holes that need to be threaded.  The pictures below show the parts for the flowers, the big down-swoops, and the up-swoops (assembled to a green acrylic disk; also showing the crystal bowl and break that will be on either side of said disk).  As you can kind of see in the pictures, the finished leaves are made of a stack of three layers and the flowers are a stack of two layers.

I assembled the flower pieces onto a couple of the light assemblies and attached those to the crystal arms.  The picture below shows those two arms hooked up to the electronics that will drive the chandelier.  Off-picture are the other four light assemblies (as of yet without flowers and not attached to their crystal arms).  Once I got everything hooked up to the driver circuitry, I fired it up – the light output is impressive!  Photons blasting out everywhere!  I don’t have any pictures of it lit up since the light would be rather blinding for the camera. 

The temperatures are surprisingly close to what my calculations said they would be.  I computed that the temperature would level out at about 60° C (140° F) when running the lights at full power.  Now that I could test the actual parts, I let it run at full power for a while.  After about an hour, the temperatures leveled out at a maximum of 58° C ± 8%. [note #1]  So anyway, woot!  The temperature is manageable and the chandelier won’t go up in flames!  (Nor will the acrylic melt, nor will the LEDs exceed their recommended operating temperatures.)

Notes:

note #1: For the curious, the tolerance I gave for the temperature readings is just that of the thermistors plus the resistors I’m using.  I could calibrate out much of the resistor variance since that’s easy to measure but I probably won’t bother since I just need an approximate reading to ensure that the temperatures aren’t getting too high.

That picture is a bit out of date – a number of design decisions have been tweaked as I’ve worked on this project.  I ended up choosing clear glass teardrop crystals instead of the red ones shown in that picture.  The arms I’m actually using are thicker and not quite as long.  The cylinder that the arms go into (and holds the electronics) will be a bit taller, and the bottom-most wooden disk will be a bit thicker. 

I’ve also changed the design of the flowers that surround the lights.  Instead of the yellow flowers and green leaves each being 3/4″ thick acrylic, they will each be made of three layers of 1/4″ acrylic with some layers being smaller than others.  The picture to the right shows a head-on rendering of my current plan for the flowers.

For the lights I’m using neutral white Luxeon® Rebel LEDs in LuxDrive’s Endor Star™ modules.  Each of these three-LED modules emits as much light as one of the halogen floodlights I have lighting my kitchen and this chandelier will have six (one in each flower), so I’m guessing I won’t be running it at full power most of the time.  And if I do my heat-sinking properly, I won’t have to replace the lights for 35 years!

My chandelier is a bit upside-down.  Most chandeliers have the lights pointing upwards (often looking kind of like candles).  I, however, have high concrete ceilings, and I don’t want to rely on light bouncing back off the ceiling, so my lights point down at the table.  This does mean, however, that unlike most chandeliers, I won’t get much light going up through all the nifty leaded crystal and acrylic above the lights.  So I’m going to wrap a bit of RGB LED strip around the top of the wood cylinder containing the electronics (the sme cylinder that the arms go into).  This will provide accent lighting for the top parts of the chandelier.

The circuit boards I designed arrived last weekend so I was able to go ahead and build the guts of the chandelier.  The drivers for the high-power LEDs and accent LEDs are controlled by a microcontroller and, like most of my creations, I threw an XBee wireless module into the mix so the chandelier will be part of my home network, controllable through my digital assistant and all that jazz.  The pictures below show the two circuit boards (the most heat-producing components are on the top board while the bottom board contains the microcontroller, the XBee module, and the drivers for the main LEDs) and my digital assistant’s control dialog for the chandelier.

I’ll finish this article with pictures of all the other cool parts and raw materials that I’ve purchased for this project.  It’s always exciting to have parts for makin’ stuff!  muwahaha!

Over the years I’ve accumulated quite a large collection of textures for use in 3D modelling on the computer.  So I had some nice cobblestone, flagstone, grass, and snow textures on hand, as well as photographs of street details like victorian-era manhole covers and street drains. I used the flagstones plus drop-shadows to depict the curbs.  I masked the snow layer with a grunge texture to allow some of the grass to show through and used a textured eraser to remove snow from key areas of the street.  The picture to the right shows one of the streets I made.

Department 56 made the buildings and other assorted props for my village, so to make the backdrop, I used pictures of some of their other buildings pasted in various orientations in layers.  Lower layers for buildings that were farther away.  For each lower layer, I shrunk the buildings a little bit more.  Between each layer is a faint blue fog.  At the very back is a texture that fades from night blue at the horizon to black at the top.  Some falling snow completed the picture.  The picture below and to the left is part of the resulting backdrop.

The next item that needed work was the streetlights.  They were intended to be powered by a pair of C-cell batteries, but as I found out last year, two C batteries could only run the streetlights for about half an hour.  Department 56 sells a power adaptor for these that can plugged in to an electrical outlet, but I figured I had enough parts on hand to make my own.  And thus began one of the most hacked up circuits I’ve ever made. 

The two C batteries were connected in series and there was nothing else in that part of the circuit, so they provided 3 volts to the lights.  The lowest voltage wall adaptor I had to spare produced 10 volts, so I would have to reduce the voltage a bit.  I had some 3.3v linear voltage regulators on hand, so I threw together a circuit with one of those plus the required capacitors.  I also threw a diode into the input to protect the voltage regulator in case I hooked up the power backwards.  I tested the output and it was indeed 3.3 volts, and hooked it up to the lights, which lit up beautifully … and then ten minutes later I burnt one finger on the voltage regulator and another on the diode.  This was running much hotter than I had expected!

The fact that the batteries could only run the lights for about 30 minutes should have been a clue – these lights draw a lot of current!  (well lots for my little bits of electronics)  It turns out that they drew 0.6 amps at 3.3 volts.  This was within spec for all the parts I used, but it was obvious that I was going to have to put a heatsink on the voltage regulator and do something about the diode as well.  Linear voltage regulators are not very efficient; they reduce the voltage by wasting the energy as heat.  Since I was going from 10 volts to 3.3 volts, this meant that 6.7v * 0.6A ≈ 4 watts was being turned into heat in the voltage regulator.  The first heatsink I tried wasn’t large enough (actually, I did some calculations later and they showed that it would keep the voltage regulator below its max allowable temperature, but it still meant the whole thing was hot enough to burn flesh, which bothered me), so I jammed that first heatsink into a larger one.  That worked quite well.  All that current was also flowing through the diode I mentioned earlier and that was also getting quite hot.  Since the diode was just there in case I made a mistake and hooked the power up backwards, I decided to just take it out of the circuit – I soldered a wire to bypass the diode.  So now I had an ugly little circuit board with wires everywhere and a diode that wasn’t actually even being used.  But it worked!

Later I found a 5v wall adaptor, so I put that to use instead of the 10v one.  This meant the voltage regulator didn’t have to reduce the voltage nearly as much, which helped with the heat problem.