new may 2012
Details of the electromagnet wiring procedure. Not shown (sic) is the spreadsheet used to design the things. 600-something turns of #30 magnet wire. Started out solenoid wound, by halfway it turned jumble-wound. No matter, shellacked paper hides the ugliness. And electrically it doesn't matter, as they're operated at basically DC anyway.
I made a jig to feed wire from the spool, directly onto the solid-iron core (DC, no eddy currents).The cores were simply chucked into a Ryobi hand drill, run at very low speed. The core was insulated with a layer of paper and shellac. Teflon blocks corralled the wire into shape, and shellac doesn't adhere to the teflon. Shellac was painted on after each winding layer. The shellac was allowed to dry, mounted in a vise, with the flying leads restrained. The somewhat delicate coil was then paper wrapped and dipped in shellac once a day for half a week. It took about a week to make the two coils (with little effort other than the winding).
This is the second test jig for the 6D4 and newly-wound magnets. (The first one was just a crap kludge where I waved magnets around by hand for proof of concept, before I began construction.) The 'scope display shows the noise voltage on the tube (eg. chaotic noisy electron current between cathode and plate), textbook grass noise.
Box construction and layout. I couldn't find a ready-made box so I constructed one from 3/8" thick Micarta with a welded steel frame. Rather overkill and a pain to assemble. Exposed metal is aluminum, with brass hardware. The Micarta is shellacked, sanded, then waxed.
Chassis layout, of even the most mundane technical gear, is an art. Especially with human-interface components, proper placement is more a matter of avoiding bad layout than worrying about corrent layout. Good design is invisible; more accurately, good layout lets the functional components speak without confusion; emphasizes what is important and leaves lesser details to be found as necessary or not at all.
The phosphor screen for the electromechanical/laser oscillograph (sic) was developed tediously, try after try. The reticle is #32 litzendraht wire (aka litz wire) stretched over glass, affixed with urethane, then coated with strontium aluminate powder. The devil is in the details. Took four attempts to get one to work; the first one took a week, the last one about four hours. There's a photograph taken at 4X magnification showing how crossed wires are treated; before the urethane has fully cured, each wire-cross is pressed down with the tip of an exacto blade. When this was not done, phosphor powder got under the crossings and made the reticle blurry.
The magic eye tube, super high tech in 1938 (still pretty today) is an awkward thing, but not hard to work with. It does get hot though. The photo here shows one "eye" open, there's actually two.
The 13-segment DIPlite display uses incandescent wire filaments. A very nice display, about twice as power-hungry as an LED display, but the physicality of the light is nicely visceral. These are not "obsolete" displays; they are still used in avionics. They are rugged, variable brightness, and readble in full daylight.
The sole input device is this simple cadmium sulphide cell. However, fancy-pants differentiation/integration (actually, a PID algorithm (proportional integral derivative)) extracts changes in light. In fact, all of the devices in this box are analog and all are driven with integral or differentiation functions, or both. This is the most math-heavy box I've made to date.
The most obvious idiosyncracy in this box is the longwave UV laser. (Most of my boxen have some bizarre anachronism). $12, post-paid, from China! Came in a lovely fake-velvet-lined black box. The case is electrically hot so it needed this plastic mount.
The most visible output device is the electromechanical oscillograph. Making optical devices is a pain in the ass; so fussy! The laser shoots up from inside the box, hits the tiny front-surfaced mirror affixed to the pretty brass thing between the legs of the horsehoe magnet of the mirror galvanometer. The fixed mirror mount is adjustable.
The mirror galvanometer -- this example probably made around 1940 but they date back to the 19th century, pretty much as it appears here -- is an extremely sensitive (nanoampere) device, a tiny coil of wire suspended on a taut wire, in a magnetic field. They're simply beautiful, working or not.
This assortment of technologies from three major eras requires a lot of interface, a lot of wiring, and lots of different voltages. 5V, 6V (mainly tube filaments), 12V (mainly analog signal processing, and fans), and 250V (tube plates aka anodes). All that power required careful airflow design, not particularly obvious here, but the plexiglas top cover will trap the heat of the two tubes as well as act as a dust collector. Air filtration was required.
The final wiring was not as messy as it appears here, in-progress. The excess length is required though, so that the top can be removed and set on it's side and not damage anything. The last picture in this table is what it looks like, "ready to go".
The trickiest part of assembling this box was the 3D puzzle made by objects attached to the top plate, rear panel and front panel, and the base of the box having to fit together when assembled. There's not a lot of extra room in there, and the tube turret sockets have high voltage on them, nesting down between 5V Arduino logic so mistakes would be costly.
The machine is starting to take it's final shape, though the airflow design isn't finalized here; I ditched the neon high-voltage lamp in the top back left corner and installed the plexiglas cover's positive ventilation/filtration system there.
