I have mounted and wired all sixty RGB LED’s under the tube support shelf. These RGB LED’s are in serial strings of 20 each, and I was able to connect them end-to-end, making the wiring much easier than the air pump motors! Each liquid filled tube in the bubble display will sit above one of the LED’s.
These LED’s are individually addressable by the Arduino that controls the entire system. In the video below I have them simply doing a simple color progression for testing.
I have all sixty of the air pumps soldered to connectors and mounted on vibration dampening foam. Thirty are in a line on top 1 and 7/8″ apart, and the other thirty motors are mounted offset just below them. The connectors will allow easy motor replacement if one should fail. (A screw, an electrical connector, and a 1/4″ tube will need to be moved over to the new air pump.)
My super high tech vibration dampening material is simply 3/8″ strips cut from donated black rubber mouse pads.
Here is the final frame (without covers except for the triangular end pieces):
I cut the extruded aluminum t-channel pieces to the correct size and assembled them with the MDF tube support and plywood end panels (not yet stained/finished).
Here is a picture of one side panel, plus a “one tube, one inch wide” bubble display that I mocked up before getting the tube support finished:
The two end panels are connected at the bottom by the tube support, which is just a MDF shelf on it’s side cut to the correct length. This piece of MDF rests on the ground along its entire length and serves as the support for the tube holder. I used a carpenters square and a level to make sure the tube support (and end panels) were square and correctly upright while attaching it. I would attach one bolt, and then square everything up before attaching the other bolt.
The tube support is screwed to the tube holder MDF using 32 wood screws. I might have been able to get away with 16…but I know 32 will hold!
As pictured here, the frame without electronics, tubes, motors or front/back covers weights in at 30.2 pounds.
Next up: Drilling 60 holes to mount the motors, and 30 or 60 holes for the motor power wires (I may combine the wires from two motors into the same hole, I haven’t decided yet) … and soldering connectors onto 60 motors (Which goes at around 20 motors per hour….).
The Tube Holder is a piece of hardwood that sits directly beneath the tubes (holding their weight) and has holes for both the air injection tubing and the RGB LED lights for each tube. I made it out of hardwood because I had to drill 120 holes in it, and because it would be screwed to the tube support upright which is MDF.
The tolerances for the holes are relatively small, so I was careful in measuring, marking, and drilling.
Every inch a 1/2″ spade bit was used to drill a hole for the air tube, and then between them a 5/16″ bit was used for the RGB LED light.
The final product looks very regular:
How do you make sixty bubble display tubes?
One. By. One. Well, actually, you do sixty duplicates of each individual step. The steps are:
I am using T-Slot aluminum extrusions for the frame of the full sized bubble display. After I made the CAD plan, the decision was made to shorten the display to 36″ tall tubes (The width will be 60 tubes / inches). This allowed me to shorten up the support legs as the center of gravity of the frame will be much lower. I haven’t actually cut any of the aluminum stock to size yet, so the photos below show pieces of aluminum extending out the back and top, and it is 72″ wide instead of 62″ wide. You can see a single tube mounted with a rubber band on the far left. The piece of cardboard on top of it is taking the place of a metal mesh that will eventually mount over the top of the tubes, running the entire width. (The bottom cross support is simply to hold the frame up, and will probably not be visible in the final assembly either.) The triangular prism at the bottom of the frame will be covered with panels of plastic or wood, and contain the electronics, LED’s and pumps.
My BubbleDisplay project needed to control sixty DC motors or solenoids to control air injection into individual columns of liquid. Due to the large number of outputs needed, I am using a chain of (74HC595) serial shift registers so that three I/O pins can control all sixty outputs. As each serial shift register has 8 outputs, this requires eight chips (for a total of 64 outputs, four are unused). The 74HC595 can not source/sync enough current to drive the motors/solenoids directly, so I am using a TO-220 N-Channel MOSFET rated at 60 volts and 32 amps (digikey: FQP30N06L-ND) to drive the load, with an 1N4001 rectifier diode to handle current spikes. Because I had to make 8 (9 for a hot spare) copies of this circuit, I decided that fabricating a printed circuit board was the only way to go.
It only took me two tries (Moving from Version 1.0 to 1.1) before I was happy with the design, which you can see (populated for testing) above. Looks a lot nicer than the prototype, right?
One morning I came back to my prototype bubble display to find this:
Obviously my final bubble display will need to have a screened mesh at the tops of the tubes to allow air to escape, but keep insects out. (But I will also need a way to re-fill the tubes, so it can’t be a permanent screen…so acrylic welded solutions are not on the table….)
After leaving a prototype bubble display running for a week (in 80-100 degree weather) I returned to find the following clear case of evaporation:
The two tubes on the outside have glycerin, while the two tubes in the center (which are several inches lower) contain water. I didn’t leave a marker of what level the tubes were filled to at the beginning of the week, but the water tubes have obviously lost several inches due to evaporation. (The continuous bubbling action helps evaporation along quite a bit…) It looks like the glycerin suffers less from evaporation, but I expect that even a glycerin filled bubble display will need periodic refills. Note that the 80-100 degree (and dry) weather may have made this problem more pronounced.
Once I had five six foot tall tubes, I wrote some code that would attempt to render text using a 5×7 font. If you can read the three letters I scroll (vertically) in the movie below I have been successful: