原创 Creating fantastic flanges

"But what are these fanciful flanges of which you speak?" I hear you cry. Well, take a deep breath, settle down in a comfy chair, and I shall expound, explicate, and elucidate (and then I'll tell you about the flanges).

This is part of my Inamorata Prognostication Engine (see also Yummy Scrummy Faceplates and Awesome Startup Sequence). Mounted on the top of the upper cabinet we find five fabulous vacuum tubes:

These really are robust little rascals -- the largest one stands a tad over 12 inches tall as illustrated below:

These are antique tubes that no longer work, but this isn't a problem because I intend to light them up using tri-colored LEDs in the form of Adafruit's NeoPixel strips. Check out this video showing an early experiment from last year.

All I'm doing in the above video is holding a strip of NeoPixels against the neck of a large tube and then sequencing the LEDs on and off, but the final result is pretty spectacular. In the real world, an optical illusion sometimes makes it appear as though the structures inside the tube are rotating in the opposite direction to the LEDs. Furthermore, the experiment above employed a strip containing only 60 LEDs per meter, but my final implementation will be based on strips boasting a phantasmagorical 144 LEDs per meter.

Now, I could simply use a hot glue gun to stick the NeoPixel strips to the bottom of the tubes as shown in option (a) toward the bottom of the sketch below, but -- let's be honest -- that's not going to happen because it would look pretty drongo. The alternative is to "wrap" each NeoPixel strip in some sort of collar, or flange, as illustrated in option (b).

Until recently, I've been toying with the idea of using a 3D printer to create these flanges, but now I'm thinking machined aluminum à la Draw2CNC might look tantalizingly tasty.

My first thought was of a simple circular flange augmented with a tubular extrusion to accommodate the cable carrying the power (+5V and GND) and data signal as illustrated below. As an aside, I'm planning on using vintage-looking cloth-covered cable -- the sort of thing you used to see on old headphones. I remember finding a supplier for this a few years ago, and I even purchased a few meters at that time, but -- sad to relate -- it's "gone to ground" somewhere in my office and -- try as I might -- I've not been able to track it down (1,000 curses).

The actual cross-section of the flange is shown toward the bottom of the above image. There would be a large channel in the middle to accommodate the NeoPixel strip, and two smaller channels at the top and bottom to accommodate rubber seals (I was thinking of using rubber bands for this purpose).

While I think about it, although the NeoPixel strip is about 2.1mm thick, the main channel in the flange needs to be deeper that this because we'll have to bend the strip between pixels. In the illustration below, 'x' = 1,000mm / 144 = ~6.95mm, which we could use to calculate the theoretical channel depth. In practice, however, we will also have to account for the way in which the material forming the strip bends, so the best way to determine the required channel depth will be to use a "suck it and see" approach (or empirical measurement, if you wish).

Even though the five vacuum tubes have different diameters, I'd like to keep things as consistent as possible -- having the same channel depths for all five flanges, for example -- which means basing the channel depth on the smallest diameter tube because this will demand the largest bending of the strip (my head hurts).

But wait, there's yet another consideration. Consider the following images of the five tubes (the "diameter" values refer to the diameter of the necks where the flanges will be located).

Back row; left-hand side; 2.61" diameter.

Back row; middle; 2.26" diameter.

Back row; right-hand side; 2.48" diameter.

Front row; left-hand side; 2.505" diameter.

Front row; right-hand-side; 2.26" diameter.

 

I'd originally planned on making simple circular flanges that could simply "slip" over the neck of the tubes. I was also thinking of giving myself say 3/64" (~1mm) clearance on either side -- that is, making the inner diameter of each flange 3/32" (~2mm) larger than the diameter of its tube's neck -- where this extra clearance would be taken up by my rubber seals.

The problem occurs with the tube in the middle of the back row; also the tube on the left-hand side of the front row. As we see from the images above, both of these little scamps have glass/metal protrusions toward the bottom of their necks. This means that we're going to have to split the flanges into two halves as illustrated below (I haven’t shown the three-channel cross-section in order to keep things simple):

Since I have to do this for two of the flanges, I might as well do it for all five for consistency. So, that's where we stand at the moment. Now I'm going to have a play with Draw2CNC to see if it will to bring a smile to my visage or a frown to my phizog. Watch this space for further developments...