How a CRT Vector Display Works

Everything a scope clock does, it does by aiming one electron beam at a glass screen coated with a powder that glows where the beam lands. That is the whole trick, and it is worth sitting with for a moment before the schematics start, because every design decision in the rest of this series — the kilovolt supply in Vol 3, the differential deflection amplifiers in Vol 4, the vector firmware in Vol 5, the tube choice in Vol 6 — falls out of the physics of a single beam in a vacuum being steered by voltages and switched on and off thousands of times a second. A cathode-ray tube is, at bottom, a controllable point of light: you can put it anywhere on the face, make it any brightness from black to glaring, and move it fast. A clock is just a list of where to put that point and how bright to make it, refreshed faster than the eye can follow. This volume takes the tube apart electrode by electrode, explains how each one earns its place, and then steps back to the question that defines the whole genre: why draw a clock face as a handful of glowing strokes instead of scanning it like a television.

The tubes this series cares about are electrostatic-deflection oscilloscope and radar CRTs — small round-faced glass envelopes, typically 1 to 5 inches across the screen, that steer the beam with voltages on internal metal plates. They were made by the millions from the 1930s through the 1970s for test gear and radar indicators, and they are exactly the wrong tube for a television (which uses magnetic deflection and a much shorter, fatter envelope). The two tubes the owner’s builds use — the Philips DG7-32 in the OSC4.4 and the RCA-family 3RP1A / 3MP1 class in the Dutchtronix world — are textbook examples, and their datasheets are the backbone of this chapter.

2.1 Anatomy of an electrostatic CRT, end to end

An electrostatic CRT is a sealed, evacuated glass envelope with a cluster of electrodes at the narrow (base) end — collectively the electron gun — a pair of deflection-plate assemblies in the neck, and a phosphor-coated faceplate at the wide end. The beam is born at the base, shaped and accelerated through the gun, steered by the plates, and slammed into the phosphor. Reading from the base toward the screen:

   BASE END                                                      SCREEN END
   (electrodes on pins)                                          (faceplate)

   Heater   Cathode   Grid     Focus     Accel    Y-plates  X-plates   Aquadag
   (filament)  K       g1      anode      anode   (D2 D2')  (D1 D1')   wall coat
      |        |       |       (g3/a1)   (a2/g2/        |        |         |
      |        |       |          |       g4/ultor)     |        |         |
   ___v________v_______v__________v_________v___________v________v_________v____
  (  HHHH )  [===]   [ () ]    [   ()   ]  [   ()   ]   == ==    == ==     )))))) ---> PHOSPHOR
   heats K   emits    gates    focuses    accelerates  deflect  deflect   final
            electrons  beam     to a spot  + collimates  vert.   horiz.   anode +
                       current             the beam      beam    beam     shield
                                                                          getter (in neck)

Each electrode in turn:

• Heater / filament. A small tungsten coil that warms the cathode to thermionic-emission temperature. It carries no signal; it just makes the cathode hot. In the tubes used here the heater runs on 6.3 V at a few hundred milliamps — the 3RP1A draws 0.6 A ± 10 %, the 3MP1 the same 6.3 V / 0.6 A, the DG7-32 is 6.3 V / 300 mA. Some smaller or European tubes run 2.5 V. The heater is the only “wear item” in normal use; running it 7 V instead of 6.3 V (common when a filament transformer is loaded lightly) measurably shortens cathode life, which is why the open-source driver in Vol 10 suggests a series resistor to trim it down.
• Cathode (K). A nickel cap over the heater, coated with an oxide mix that boils off electrons when hot. It is the source of the entire beam. In a “unipotential cathode” tube (the kind here) the cathode is electrically separate from the heater, so the cathode can sit at the most-negative potential in the gun while the heater floats near it (the datasheets cap heater-to-cathode voltage at ±125 V to avoid arcing through the thin insulation).
• Control grid (g1, the Wehnelt / brightness electrode). A metal cup with a small aperture, surrounding the cathode. Making g1 more negative than the cathode repels emitted electrons and throttles the beam current — this is the brightness control, and it is the same knob the firmware twiddles to blank the beam (see 2.2). Drive g1 negative enough and the beam shuts off entirely; the datasheets call this visual extinction or cutoff. For the 3RP1A, cutoff is about −6.75 % of the final anode voltage (≈ −67.5 V at 1000 V ultor); for the DG7-32, g1 cutoff is −50 to −100 V.
• Focus anode (g3 / a1). A cylindrical electrode at a moderate positive voltage that, with the surrounding electrodes, forms an electrostatic lens — a region of curved equipotential surfaces that bends the diverging electron paths back to a fine crossover at the screen, exactly the way a glass lens focuses light. Adjusting its voltage (the FOCUS pot) changes the lens “power” and brings the spot to a sharp point. The 3RP1A wants its focus (g3) electrode at roughly 16.5–31 % of the final anode voltage; the DG7-32 focus (g3) runs 0–120 V over an accelerator of a few hundred volts.
• Accelerating anode (a2 / g2 / g4 / “ultor”). The highest-voltage electrode in the tube, held at the full anode/ultor voltage. It accelerates the now-focused beam to its final velocity so it hits the phosphor hard enough to glow. In oscilloscope-tube data this final electrode is called the ultor (RCA’s term) and is often the same node as g2/g4 and the internal collector, tied together inside the glass. The 3RP1A and 3MP1 are rated for 2500 V max on this electrode; the DG7-32, a deliberately low-voltage tube, runs 400–800 V.
• Deflection plates (two pairs: X and Y). Two pairs of flat plates straddling the beam in the neck. A voltage difference across a pair pushes the beam sideways. One pair deflects vertically (Y), the other horizontally (X). These are the signal electrodes — the deflection amplifiers of Vol 4 drive them — and they are the heart of why this tube is easy to use as a vector display (2.3). In RCA notation the four plate terminals are DJ1–DJ4; DJ1/DJ2 are nearer the screen, DJ3/DJ4 nearer the base, and the two pairs are mounted at 90° ± 3°.
• Aquadag coating. A conductive graphite (“Aquadag”) layer painted on the inside of the flared glass between the plates and the screen, connected to the final anode. It does two jobs: it continues accelerating/guiding the beam after the plates (a field-free drift region), and it collects the secondary electrons knocked loose when the beam hits the phosphor, giving them a path back to the supply so the screen doesn’t charge up and repel the beam. The DG7-32 datasheet notes a transparent conductive layer between glass and phosphor tied to the accelerator for the same reason — and as a bonus it lets you run the tube with the screen near high potential without your finger on the glass distorting the picture.
• Phosphor screen. The powder coating on the inside of the faceplate that converts beam energy to light (2.4). Its type (P1, P7, P31…) sets the colour and persistence.
• The getter. A small ring or cup of reactive metal (barium is typical) flashed during manufacture to a silvery mirror on the inside of the neck or bulb. It chemically mops up the last traces of gas to keep the vacuum hard for the tube’s life. A getter that has turned white/milky is the classic sign of a tube that has gone soft (lost vacuum) and is scrap — a point that matters when buying surplus tubes in Vol 6.

2.1.1 Reading a real datasheet — the 3RP1A pinout

The abstract anatomy maps directly onto pins. The RCA 3RP1A (a 3-inch precision oscillograph tube, “tentative data” July 1955) shows the whole family in one table:

  Pin   Electrode                     Role
  ----  ----------------------------  -------------------------------------------
   1    Heater                        6.3 V filament
   2    Grid No.1 (g1)                brightness / blanking (Wehnelt)
   3    Cathode (K)                   beam source
   4    Grid No.3 (g3)                focus anode
   8    Ultor (g2 + g4 + collector)   final accelerator, up to 2500 V
   9    Deflecting electrode DJ2      \  one deflection pair
  10    Deflecting electrode DJ1      /  (DJ1/DJ2 nearer screen)
   6    Deflecting electrode DJ3      \  other deflection pair
   7    Deflecting electrode DJ4      /  (DJ3/DJ4 nearer base)
  12    Heater                        6.3 V filament (other end)
  5,11  Internal — Do Not Use         (omitted on the 10-pin base)

Note how few terminals there are: a heater, a cathode, two grids, one high-voltage anode, and four deflection-plate connections. There is no signal path that is not either DC bias or a deflection plate — which is exactly why a scope clock’s electronics divide so cleanly into “a power supply that makes the bias rails” and “amplifiers that wiggle the plates and the grid.”

2.2 The electron gun and beam formation

The gun is the cathode plus the grid plus the focus and accelerating anodes, and its job is to turn “a hot cathode emitting electrons in all directions” into “a sharp, bright, controllable spot on the screen.” Three things happen in sequence.

First, emission. The heated cathode emits a cloud of electrons. Left alone they would just sit there as a space charge.

Second, current control at the grid. The control grid (g1) sits just in front of the cathode with a small hole. Its voltage relative to the cathode sets how many electrons get through the hole and into the rest of the gun — i.e. the beam current, which is the brightness. This is a classic triode-style transfer characteristic: with g1 a little negative, some current flows; make g1 more negative and the current falls steeply; past cutoff, nothing gets through and the screen is dark. The 3RP1A’s “average characteristics” curve shows beam (ultor) current and screen brightness both rising sharply as g1 goes from about −45 V toward 0 V at a 1000 V ultor — a roughly square-law knee, not a straight line, which is why brightness control feels nonlinear and why the video amplifier in Vol 4 has to swing a fair number of volts to go from “off” to “full bright.”

Two ways to drive that control voltage exist, and both scope-clock camps use one of them:

• Grid-drive. Hold the cathode at a fixed (negative) bias and drive g1 with the video signal. Conceptually clean, but the grid circuit ends up awkward to bias.
• Cathode-drive. Hold g1 at an adjustable brightness voltage (a pot) and drive the cathode with the video signal — pulling the cathode down brightens the beam, pushing it up blanks it. This is what TubeTime’s open-source driver does: “drive the grid to ground (or… an adjustable brightness voltage from a potentiometer) and connect the cathode to the amplifier output. The end result is the same.” There is a slight side-effect on the focus electrode, which the focus pot trims out.

Either way, the key fact for a clock is this: the beam is off by default and is only switched on (unblanked) where the firmware wants a line or dot. The Z-axis / video signal that does this switching is covered as a circuit in Vol 4 and as firmware in Vol 5; here it is enough to know it acts on g1-vs-cathode.

Third, focusing and acceleration. The diverging beam passes through the focus anode, whose field acts as an electrostatic lens, and is brought to a fine crossover at the screen plane. The accelerating anode (ultor) then gives the beam its final energy. Brightness depends on both beam current (grid) and accelerating voltage: more kV makes the same number of electrons hit harder and glow brighter. The 3RP1A datasheet is explicit that “brilliance and definition decrease with decreasing ultor voltage,” recommending against running below ~500 V except in dim rooms — a direct argument for the stiff kilovolt supply of Vol 3, and for post-deflection acceleration (2.5) on bigger tubes.

A subtlety the datasheets flag: focus voltage, cutoff voltage, and deflection sensitivity all scale with the accelerating voltage. The 3RP1A lists focus as a percentage of ultor (16.5–31 %) and cutoff as −6.75 % of ultor precisely because the gun’s geometry is fixed and only the voltages scale. Change your anode rail and you must re-trim focus and re-scale your deflection drive. This coupling is why calibration (Vol 12) is done in a fixed order: set anode, then focus, then size.

2.3 Electrostatic vs magnetic deflection

There are two ways to bend an electron beam: an electric field between charged plates (electrostatic), or a magnetic field from current in a coil (magnetic). Televisions and computer monitors used magnetic deflection — a yoke of coils slipped over the neck — because it can bend a high-energy beam through the wide angles a big short-necked picture tube needs, and the coils sit outside the glass. Oscilloscopes and radar indicators used electrostatic deflection, and so do scope clocks, for reasons that are decisive for a maker:

• It is trivially easy to drive from an amplifier. A deflection plate is, electrically, just a small capacitor (a couple of picofarads — the 3RP1A lists ~2 pF plate-to-plate, the DG7-32 1.0–1.7 pF). To move the beam you set a voltage, and a voltage amplifier with a high-value load resistor does that directly. A magnetic yoke needs a current amplifier driving an inductor, with flyback and energy-recovery complications — a whole different and harder design.
• It is fast. With only picofarads of plate capacitance to charge, the beam responds to voltage changes essentially instantly; the bandwidth limit is the amplifier, not the tube. That is exactly what you want for drawing crisp vectors and for the fast blanking edges between strokes.
• The whole steering mechanism is inside the sealed tube, on dedicated pins, so there is nothing mechanical to align.

The cost is deflection sensitivity: electrostatic plates need a meaningful voltage swing to move the beam across the face, and the higher the accelerating voltage, the stiffer the beam and the more volts per centimetre you need. Datasheets quote this as a deflection factor — volts required per unit of beam displacement, usually normalized per kV of anode voltage:

  Tube       Plate pair   Deflection factor                  Final anode
  --------   ----------   -------------------------------    -----------
  3RP1A      DJ1/DJ2      73–99  V dc / inch / kV of ultor    up to 2500 V
             DJ3/DJ4      52–70  V dc / inch / kV of ultor
  3MP1       DJ1/DJ2      115–145 V dc / inch / kV of anode   up to 2500 V
             DJ3/DJ4      110–140 V dc / inch / kV of anode
  DG7-32     Y axis (N1)  0.35–0.43 mm / V  (≈ deflection)    400–800 V
             X axis (N2)  0.24–0.30 mm / V                    (typ. 500 V)

Worked example: the 3RP1A at a 1000 V ultor needs about 73–99 V dc per inch on DJ1/DJ2 and 52–70 V per inch on DJ3/DJ4. To swing the spot across a ~2-inch useful screen you therefore need on the order of ±100–200 V differential on each pair — which is precisely why the small scope-clock tubes are driven with plate signals centred in the 150–250 V region. The DG7-32, running at only 500 V, is far more sensitive (its datasheet quotes deflection in mm per volt, ~0.4 mm/V vertical), which is part of why low-voltage tubes are friendly for simple builds.

2.3.1 Why each pair is driven push-pull (differentially)

You will never see one deflection plate grounded and the other driven, even though that would bend the beam. Both plates of a pair are driven antiphase about a common centre voltage — as one goes up the other comes down by the same amount — for two reasons:

1. It keeps the average plate voltage constant, so the mean potential the beam sees in the deflection region doesn’t shift as you steer. Single-ended drive changes the average field, which defocuses and shifts the spot (the “spot defocusing” the 3RP1A datasheet warns about, recommending the ultor be returned to a point giving the lowest potential difference to the plates). Differential drive cancels that.
2. It halves the swing each amplifier must produce and keeps the deflection linear and symmetric about screen centre.

So each axis needs a complementary pair of high-voltage drive signals centred on a bias in the mid-hundreds of volts — exactly the differential cascode amplifiers TubeTime describes (“the core of the differential amplifier is a matched dual NPN transistor… a cascode stage… level shifts the signal up to about 1 kV”) and the subject of Vol 4. The astigmatism control found on better designs trims the final-anode voltage relative to the average plate voltage to keep the spot round rather than oval — a direct consequence of getting that average exactly right.

2.4 Phosphors and persistence

The phosphor is the powder on the faceplate that turns the beam’s kinetic energy into visible light. Two properties matter for a clock: the colour (where the emission peaks) and the persistence (how long it keeps glowing after the beam moves on). Persistence is the lever that trades off against flicker and refresh rate, and choosing it well is most of what makes a scope clock look good rather than strobey.

Phosphors are catalogued by a “P-number” (the RMA/EIA scheme). The ones relevant to scope clocks:

  P#    Colour              Persistence        Where you meet it / notes
  ----  ------------------  -----------------   ----------------------------------------
  P1    green (~525 nm)     medium (~ms)        the classic scope green; 3RP1, 3MP1, DG7-32
  P2    blue-green          long (>1 min dim)   radar; too laggy for a moving clock face
  P7    blue-white flash +  cascade: short      long-persistence radar; bright blue flash
        yellow-green glow   blue, LONG yellow    then a slow yellow-green afterglow
  P31   yellow-green (GH)   medium-short        bright, efficient scope green; very common
  P39   green (~525 nm)     long                Zn2SiO4:Mn,As — low-flicker display tubes
  P11   blue (~460 nm)      short (~2 ms)       photographic scope work; cold blue

A few practical readings of that table:

• P1 is the default “oscilloscope green” and is what the owner’s tubes carry — the 3RP1A, 3MP1, and DG7-32 datasheets all say green fluorescence, medium persistence, peaking around 525 nm where the eye is most sensitive. It is the safe, period-correct choice for a clock and is bright enough at the few-hundred-volt to ~1 kV anode voltages these designs run.
• P31 is a brighter, efficient yellow-green that many modern builders prefer; it refreshes cleanly and is forgiving on a low-voltage supply.
• P7 is the long-persistence radar phosphor. It is a cascade: a short blue flash where the beam currently is, on top of a slow yellow-green afterglow that lingers for seconds. On a clock this gives a ghostly trailing look — atmospheric, but it smears motion (a sweeping seconds hand becomes a comet tail) and is prone to burn-in. Use it for effect, not for a crisp face.
• Burn-in is real. As the NYC Resistor write-up warns, a stationary bright spot — e.g. the bottom-left corner when the DACs are tri-stated during firmware flashing — will permanently dim that patch of phosphor. Keep the beam moving, blank it when idle, and turn brightness down while reprogramming.

The persistence-vs-refresh trade is the core design tension. The screen is being repainted, not held: each stroke glows, fades, and must be redrawn before it fades too far or the eye sees flicker. With a medium-persistence phosphor like P1/P31 the firmware must refresh the whole face at roughly 40–60 Hz or faster to look steady — the same threshold as film and CRT TV. Draw a complex face too slowly and it flickers; draw a simple one and you have refresh time to spare for extra detail. A long-persistence phosphor (P7, P39) lets you refresh slower without flicker, at the cost of smearing any motion and risking burn-in. For a clock with a moving second indicator, medium persistence plus a fast refresh wins. (How the firmware budgets that refresh time across strokes is Vol 5; the brightness/contrast of a single stroke also depends on how long the beam dwells on it — the dwell-time story in 2.6.)

FIGURE SLOT 2.4 — Side-by-side of the same clock face on a P1/P31 green medium-persistence tube vs a P7 long-persistence tube, showing the trailing afterglow on P7. Suggested: build photos once both phosphor types are on the bench, or Openverse scope-screen photos.

2.5 Post-Deflection Acceleration (PDA)

There is a fundamental tension in 2.3 and 2.4: high accelerating voltage means a bright, crisp spot (2.4) — but it also makes the beam stiff, so deflection needs more volts and the picture gets smaller and harder to drive (2.3). For small tubes at ~1 kV you can have both. For large or very bright tubes you cannot, and the elegant fix is post-deflection acceleration.

The idea: keep the beam relatively slow while it passes between the deflection plates (so it deflects easily with modest plate voltage), then accelerate it hard after it has been steered, on its way to the screen. A separate, much higher-voltage electrode — a spiral or band of conductive coating near the screen end, the PDA or “intensifier” anode — does the post-steering acceleration. The beam gets its brightness without the plates having to fight a stiff beam.

PDA’s cost is voltage. The post-deflection anode wants to run well above the main accelerator — into the multi-kilovolt range — and the usual way to make that from a modest main supply is a voltage multiplier (Cockcroft-Walton style: diodes and capacitors that rectify and stack an AC waveform up to several times its peak). TubeTime’s design makes this concrete: tap the 1 kV supply’s diode, add a “2-stage multiplier… can generate 3 kV which is fine for 3″ and 5″ tubes that require it.” Larger tubes (and the high-brightness clear-faced tubes some premium builds favour) need this; the small tubes in the owner’s two builds mostly do not. The generation and taming of these multiplied-kV rails is the subject of Vol 3, and the amplifier consequences (the plate drive can stay low because the beam is slow during deflection) are revisited in Vol 6 when choosing a tube.

2.6 Vector vs raster: why a clock is drawn as strokes

Here is the conceptual heart of the whole genre. A television draws a raster: the beam scans a fixed grid of lines, left-to-right and top-to-bottom, every frame, and the picture is made by modulating brightness as it sweeps — the beam visits every point whether or not anything is there. A scope clock draws a vector image: the beam is steered directly to the points and lines that make up the figure and goes nowhere else. The face is described as a list of strokes — “pen down here, draw to there, pen up, jump to the next stroke” — exactly like a pen plotter, but with an electron beam that has no inertia.

For a sparse image like a clock face — a circle, twelve tick marks, two or three hands, maybe a digital readout — vector wins decisively:

• Bandwidth. A raster of even modest resolution forces the beam through every pixel every frame, so the brightness signal must change at pixel rate: tens of MHz of video bandwidth for a sharp picture. A vector clock only has to move the beam between the few dozen endpoints that define the face, plus draw the connecting strokes. The information content is a handful of coordinates, not a full bitmap. The deflection amplifiers in these designs run only ~10–15 kHz bandwidth (TubeTime’s figure) and still draw a crisp face, because they only ever trace the lines that exist. The video (Z) amplifier needs more speed (TubeTime’s runs ~6 MHz) — but only to switch the beam cleanly on and off at stroke boundaries, not to paint a full raster.
• Crispness. A vector line is a continuous glowing stroke at full brightness along its whole length — no pixel grid, no stair-stepping from the display itself (any “jaggies” come from the DAC stepping, not the tube). The hairline-sharp hands are why a scope clock looks like an instrument, not a tiny television.

2.6.1 Blanking, the Z axis, and beam-settling

Two artifacts make the difference between a clean vector image and a messy one, and both come straight from the physics of 2.2 and 2.3.

Retrace must be blanked (the Z axis). When the beam jumps from the end of one stroke to the start of the next, it sweeps across the screen and would draw an unwanted line connecting them — unless you switch it off during the jump. That switching is the Z-axis (blanking) signal acting on the grid/cathode (2.2). The NYC Resistor demo, which had no Z-axis control, shows the problem perfectly: the photos have visible “retrace” lines linking the digits because the beam stayed on while moving between characters. A real scope clock blanks during every retrace, so you see only the strokes that are meant to be there. The firmware emits an X, a Y, and a Z value for every move (Vol 5).

Beam-settling time is finite. Even though the beam itself is essentially massless, the voltages driving the plates do not change instantaneously — the amplifier has to charge the plate capacitance through its load resistor, and it has finite slew rate. So after commanding a jump to a new point, you must wait a few microseconds for the deflection voltages to settle before unblanking, or the start of the stroke lands in the wrong place and you get a faint smear leading into it. And because brightness depends on dwell time — a spot the beam lingers on glows brighter than one it crosses quickly — the bright dots at the ends of strokes seen in the NYC Resistor images come from the beam pausing there while the firmware computes the next coordinate. Good vector firmware equalizes dwell (sometimes padding with NOPs so every segment takes the same time) so the line looks uniform. Settling time, dwell equalization, and blanking timing are the fiddly heart of vector firmware in Vol 5.

  One stroke, in the time domain:

   Z (beam):  ____|‾‾‾‾‾‾‾‾‾‾‾‾‾|________|‾‾‾‾‾‾‾‾‾  on while drawing, off while jumping
                  ^settle       ^blank   ^settle
   X,Y:       jump→ ramp along stroke → jump → ramp ...
                    (beam ON here)      (beam OFF here)

2.7 The heritage: drawing with a beam

The scope clock did not appear from nowhere; it sits at the end of a long line of “draw-with-the-beam” devices, and knowing the lineage explains why the aesthetic is so deliberately retro.

It begins with the oscilloscope itself. Put a sine wave into one axis and another into the perpendicular axis and the spot traces a Lissajous figure — the looping curves that are the original X-Y drawing on a CRT, and the direct ancestor of every scope-clock arc. Vol 1 already made the comparison; the genre is literally “an oscilloscope in X-Y mode, told what to draw by a microcontroller.” Premium clocks lean into this: the Oscilloclock “OscilloTerm” renders all its text and graphics “using beautiful, curvy Lissajous figures” via a sine-cosine drawing scheme.

Next came radar. The PPI (plan-position indicator) is a vector display: a radial sweep rotates in sync with the antenna and the beam brightens on echoes, painted on a long-persistence phosphor (P7, P2) so the picture lingers between sweeps. That is where long-persistence phosphors and the whole vocabulary of “intensity modulation” and “afterglow” come from — the same phosphor table in 2.4 that a clock builder reads today.

Then vector arcade and home games. Through the late 1970s and 1980s, before raster could do crisp line graphics cheaply, a wave of games drew their worlds as glowing vectors: Atari’s Asteroids and Tempest, Cinematronics titles, and the home Vectrex console (a built-in vector CRT). These used analog “vector generators” — integrators that ramp the deflection voltages smoothly between endpoints — to get truly continuous lines, the technique commenters on the NYC Resistor piece contrast with the microcontroller’s stepped Bresenham approximation. The through-line to a scope clock is exact: same tube physics, same blanked-retrace discipline, same “a list of strokes” data model. TubeTime’s open-source driver was literally built for Asteroids and Flappy Bird arcade reproductions before it was a clock board.

Finally, the playful modern end: people draw anything on a scope. The NYC Resistor project put a 24-hour analog clock, an N-body planetary simulation, scrolling tweets, and an Asteroids-style game on a junk-box oscilloscope with a Teensy and two R-2R DAC ladders. And the Oscilloclock “OscilloTerm” goes furthest of all — a VT52-compatible serial terminal rendered in Lissajous strokes that can run a Zork session piped from a dfrotz interpreter, displaying Infocom’s 1981 copyright notice in curvy green text on a clear B7S4 tube at 2.1 kV. A clock that can also play Zork is the whole spirit of the thing: once you can put the beam anywhere and switch it on and off fast enough, the clock face is just the default screen.

The rest of the engineering series builds outward from this chapter: Vol 3 makes the anode, focus, bias, and heater rails that bias every electrode named in 2.1; Vol 4 builds the differential deflection amplifiers of 2.3 and the video/Z amplifier of 2.2/2.6; Vol 5 writes the firmware that emits the X/Y/Z stroke lists of 2.6; and Vol 6 helps you pick a tube whose phosphor, voltage, and deflection factor (2.4, 2.3, 2.5) match the supply and amplifiers you intend to build.

2.8 References (Vol 2)

• RCA, “3RP1-A Oscillograph Tube” tentative data sheet, July 1955 (electrostatic focus & deflection; pinout, ultor/focus/cutoff voltages, deflection factors, typical oscillograph circuit). Held in 02-inputs/CRT Data Sheets/3RP1A-rca.pdf.
• RCA, “3MP1 Oscillograph Tube” tentative data sheet, July 1950 (anode-No.2/grid-No.2 = “ultor”, deflection factors, cutoff). Held in 02-inputs/CRT Data Sheets/3MP1.pdf.
• Philips, “DG7-32 — Low voltage cathode ray tube for oscilloscopes” data sheet, 1957–59 (6.3 V / 300 mA heater, 400–800 V accelerator, g1 cutoff −50 to −100 V, deflection in mm/V, conductive inner coating). Held in 02-inputs/OSC4_4 (I have this)/DG7-32.pdf.
• E. Schlaepfer (TubeTime), “Electrostatic CRT Driver Design,” tubetime.us (differential cascode deflection amps, ~10–15 kHz; ~6 MHz video amp; cathode-drive vs grid-drive; PDA via voltage multiplier; focus/astigmatism trim; 6.3 V filament caution). Held in 02-inputs/A - Open Source/TubeTime » Blog Archive » Electrostatic CRT Driver Design.pdf.
• T. Hudson / NYC Resistor, “Vector Display Introduction,” nycresistor.com, 2012 (R-2R DAC X-Y drive, Bresenham line/Hershey-font strokes, retrace/Z-axis blanking, dwell-time bright dots, burn-in). Held in 02-inputs/PDF Resources, Ideas, and Schematics/Vector Display Introduction » NYC Resistor.pdf.
• P. Jankowiak (KD5OEI), “Cathode Ray Tube Phosphors — Of Interest To The Experimenter,” rev. 2010 (P-number colours, persistence, compositions; P1/P7/P31/P39 data). Held in 02-inputs/Electrostatic CRT Project/www.labguysworld.com/crt_phosphor_research.pdf.
• Oscilloclock.com, “Zork on an OscilloTerm!” (VT52 terminal rendered in Lissajous figures, Zork via dfrotz, 2.1 kV clear B7S4 tube). Held in 02-inputs/PDF Resources, Ideas, and Schematics/Zork.pdf.
• “Oscilloscope Clocks” resource compilations. Held in 02-inputs/PDF Resources, Ideas, and Schematics/.
• General background: oscilloscope X-Y mode and Lissajous figures; radar PPI displays; Atari vector arcade hardware and the GCE Vectrex; Cockcroft-Walton voltage multipliers.