Safety, Calibration, Troubleshooting & Cheatsheet

This is the volume that the whole series has been deferring to. Every path in Vol 1’s decision tree, every supply in Vol 3, every amplifier in Vol 4, and every build in Vols 7 through 10 ends at the same place: a glass tube with a few hundred volts on its plates, a near-three-hundred-volt rail behind a reservoir capacitor that does not politely discharge itself when you pull the plug, and an evacuated envelope that can throw glass across a room if you crack it. The good news is that the discipline that keeps you alive around a scope clock is small, learnable, and the same every time. The bad news is that it is not optional, and the failure mode is not “the magic smoke escapes” — it is a hand-to-hand current path across your heart. Read this volume before you energize anything, keep the cheatsheet in section 12.6 within arm’s reach of the bench, and treat the “before every power-up” checklist as a pre-flight ritual rather than a suggestion.

12.1 Safety — the mandatory brief

The hub-wide baseline lives in the shared safety note (_shared/safety.md), which classes the Scope/CRT family as the highest-voltage tier in the entire Clocks project — above the Nixie builds, well above anything mains-only. This section builds the scope-specific discipline on top of that baseline. If you read nothing else in this volume, read this.

12.1.1 The hazard stack

A scope clock stacks several independent hazards, and the discipline below addresses each. Understand what you are dealing with before you reach for a screwdriver.

• The accelerating / post-deflection-acceleration (PDA) anode supply. Small electrostatic tubes of the kind these clocks use run their final anode (A2 / aquadag) anywhere from a few hundred volts to several kilovolts. The two owned builds are gentle members of the family — the OSC4.4 develops roughly 300 V DC on its primary rail and accelerates the beam from a negative cathode rather than a high positive anode (see the test-point table in 12.5) — but the larger tubes covered in Vol 6 and the stiff open-source supply in Vol 10 reach true kilovolt territory. Treat any anode/PDA connection as lethal until proven otherwise.
• Line-derived rails around 300 V. The OSC4.4 high-voltage section rectifies a transformer secondary to about 300 V DC (verified at C6−/R8+ during bring-up). Some designs — and many donor oscilloscope chassis you might cannibalise — derive their rails directly from the mains with no isolating transformer at all. Those are the most dangerous of all because the chassis “ground” can sit at full line potential.
• Stored charge after power-off. Reservoir and filter capacitors hold their charge for seconds to minutes after the plug is pulled, and a CRT’s own internal capacitance (the aquadag coating against the outside ground coating) behaves like a capacitor that can hold a charge for hours. Unplugging the clock does not make it safe. This is the single most common way experienced people get bitten.
• CRT implosion. The tube is an evacuated glass envelope. A crack or a sharp knock, especially near the thin neck or the flat faceplate, can collapse it inward and then spray glass outward. The bigger the tube, the more energy is stored in that vacuum.

12.1.2 The discharge procedure

Never trust a “they probably bled down by now.” Discharge actively, then verify.

1. Power down and unplug at the wall. Do not rely on a switch.
2. Wait, but do not rely on waiting alone — go straight to the bleeder.
3. Discharge through a bleeder resistor, never a bare screwdriver. A bare screwdriver across a charged HV cap produces a bang, a pit in the blade, a welded contact, and a current spike that can damage the cap and startle you into the chassis. Use a discharge tool: a high-value, suitably rated resistor (a 1–10 MΩ, 2 W-plus part, or a purpose-built HV discharge probe) with one end clipped to chassis ground and the other touched to the node, held there for several seconds so the RC settles. For the CRT anode/aquadag, clip the ground lead to the tube’s external ground coating (the dag) and discharge the anode cap/connection through the resistor.
4. Verify 0 V with an HV-rated meter. Confirm the discharge actually happened — measure the node against chassis ground with a DMM and probes rated comfortably above the working voltage. Do not assume; measure. Only a verified 0 V reading clears a node for contact.
5. Re-verify after any pause. Some capacitors exhibit dielectric absorption — they recover a surprising fraction of their charge minutes after you discharge them. If you walked away, discharge and verify again before touching.

12.1.3 The standing rules

These are non-negotiable around any energized scope clock, and they are the same rules the shared safety baseline states for every HV build in the hub.

• One hand behind your back. When you must probe a live HV node, keep one hand in your pocket or behind your back. This keeps current off the hand-to-hand path that runs across your heart. A shock through one hand to a foot is survivable far more often than one across the chest.
• Isolation transformer for anything line-derived or transformerless. Put a bench isolation transformer between the wall and the clock whenever the supply is mains-derived, and always when working on a transformerless chassis. It breaks the path to earth that turns a single touch into a circuit.
• Permanent bleeder resistors stay fitted. The finished build should self-discharge. Fit a permanent bleeder across the main HV rail so the reservoir bleeds down within a reasonable time after power-off. This protects future-you and anyone who opens the case.
• HV-rated tools only. Meter, probes, and leads rated well above the working voltage; insulated-handle tools; no cheap pocket DMM on a kilovolt node.
• Eye protection and tube-handling care. Wear eye protection whenever a bare CRT is out of its enclosure. Handle the tube by the wide body, never by the fragile neck; do not knock the faceplate or neck; store and transport it protected. Respect the implosion risk every time the glass is exposed.
• Never work alone on a live kV supply. Have someone within earshot who knows where the plug is and how to kill the power. If the supply is in the kilovolt class, this is a hard rule, not a preference.
• Respect creepage and clearance. Keep HV traces, wires, and terminals spaced for the voltage they carry; do not let a tight steampunk enclosure (Vol 11) force conductors closer than the working voltage allows. Contamination, humidity, and dust all reduce the spacing you can get away with — leave margin.

12.1.4 Before every power-up — the checklist

Run this every time, even on a board you energized five minutes ago. Habit is what saves you on the day you are tired and distracted.

1. No conductive tools, solder offcuts, or stray wire clippings on or under the board.
2. CRT pins wired to the correct board terminals (re-check K, G, F, A, the four plates, and the two filament pins against the pinout in 12.5 — a swapped K/G is a dead spot at best).
3. Bleeder resistor present and connected across the HV rail.
4. Filament dropping resistor (Rfilament) is the right value for your tube; you have a plan to measure filament voltage on first power-up.
5. Isolation transformer in line if the supply is mains-derived or transformerless.
6. HV-rated DMM set, leads in the correct jacks, ground clip on chassis/dag.
7. One hand rule in force the moment power is applied; the other hand stays away.
8. Eye protection on if the tube is exposed.
9. You know where the wall plug is and can reach it without leaning over the chassis.
10. A second person is within earshot if the supply is in the kV class.

12.2 Calibration — the bring-up and adjustment procedure

Calibration of an electrostatic scope clock is the act of coaxing the beam into a single round, sharp, correctly-placed, correctly-sized spot — and then trusting the firmware to turn that spot into a clock face. The OSC4.4 assembly instructions give an unusually clean, staged bring-up that doubles as a teaching example for any of these designs, because the adjustments map one-to-one onto the universal CRT controls discussed in Vols 2 and 4. This section generalises that procedure and explains what each pot physically does.

12.2.1 What each adjustment physically changes

Before turning anything, understand the mechanism. The CRT gun (Vol 2) is a stack of electrodes: cathode (K), control grid / Wehnelt (G), focus electrode (F), and the accelerating anode (A). Each pot acts on one of these or on the deflection plates.

Pot (OSC4.4)	Controls	What it physically changes
P1	Focus voltage (F)	Sets the voltage on the focus electrode, squeezing the electron stream to a fine crossover at the screen. Too far either way blurs the spot into a disc.
P2	Anode / spot roundness (A)	Trims the anode voltage feeding the spot-shaping/astigmatism balance; turned for the roundest, sharpest spot once focus is close. Interacts with P1.
P3	Brightness (grid bias)	Makes the grid more or less negative relative to the cathode, gating how many electrons get through — the intensity / Z-axis bias. Also subtly affects achievable focus.
P4	Image size, one axis	Scales the deflection drive on its axis (gain of the deflection amp), setting how far a given firmware coordinate throws the beam.
P5	Image size, other axis	As P4, the orthogonal axis. P4 and P5 together set overall picture size and aspect.
P6	X centring (side adjust)	Adds a DC offset to the X plate pair, sliding the whole image left/right.
P7	Y centring (side adjust)	Adds a DC offset to the Y plate pair, sliding the whole image up/down.

On the Dutchtronix unit the philosophy is the same but the knob is different: that clock drives an external oscilloscope in X-Y mode, so its single “size” control is R8 (turn fully counter-clockwise for maximum image, then clockwise to fit the host scope’s graticule), and focus/brightness/centring belong to the host scope’s own front panel. The Dutchtronix operating notes have you set the scope to X-Y, 0.5 V/division, DC-coupled, and trim R8 (and the host’s VAR) until the face squares up to eight divisions. The concepts in the table above still apply — they have just moved onto the oscilloscope’s controls.

12.2.2 The OSC4.4 bring-up sequence (worked example)

This is the staged procedure from the OSC4.4 assembly instructions, with the why added. The key idea is get a spot first with the brains removed, then add intelligence one IC at a time so that if something breaks you know exactly which stage did it.

Stage 0 — preset the controls. Set P1 through P5 to their mid positions. Set P7 (side adjust) fully counter-clockwise and P6 (side adjust) fully clockwise. These extremes deliberately park the beam so that the forthcoming “blob” lands somewhere findable rather than off-screen.

Stage 1 — force a static spot with jumpers. With the MCU and DAC ICs not yet inserted, jumper U1 pin 28 to pin 20 and U1 pin 8 to pin 26 using short wire links that seat snugly in the empty socket. These links tie the deflection inputs to fixed levels so the beam parks as a stationary blob instead of being steered by absent firmware. Apply AC power; after a couple of minutes’ warm-up a blob or dot should appear near screen centre.

Stage 2 — focus and shape the spot. Adjust P1 until the spot is as round as it will go, then P2 until it is as sharp as it will go; nudge P3 (brightness) if it helps you judge focus, then go back and re-touch P1. You are converging on a clean, round, sharp dot. This is the single most important calibration step — everything downstream assumes a good spot. Unplug when satisfied.

Stage 3 — add the deflection DACs. Insert the two ICs marked “4132” into sockets U5 and U4, notch oriented up / toward the board centre. Re-power. Bring P6 and P7 back to mid to recover the spot, centre it, and re-touch P1/P2 for best focus. You now have a centrable, focusable spot driven through the deflection path.

Stage 4 — add the shifter. Unplug, insert the 12F629 into U6, re-power. When the spot reappears, watch for it to twitch slightly left/right and up/down — that confirms the shifter (the small helper that nudges the image to even phosphor wear) is alive.

Stage 5 — add the brains and draw the face. Unplug. Insert the 7528 (20-pin DAC) into U3 and the big 18F26K20 MCU into U1, both correctly oriented. Re-power and press S2 once: the clock face should appear. Now set size with P4 and P5 to taste, then make one final pass on P1/P2 for the sharpest face. Only after the face is good should you fit the optional Wi-Fi or GPS module — bring the tube fully to life first, as the instructions stress.

FIGURE SLOT 12.2 — Photo sequence of the OSC4.4 bring-up: (a) the jumpered blob, (b) a focused static spot, (c) the first clock face after S2. Build photos to be taken on the owned unit.

12.2.3 Generalising the adjustments

Every electrostatic scope clock exposes some subset of these controls; the names differ but the physics is the one taught in Vols 2 and 4.

• Focus sets the spot’s fineness (the crossover at the screen). Always converge focus on a static spot or a slow figure, never on a busy face.
• Astigmatism / spot roundness corrects an oval spot — a spot that focuses sharp in one axis and fuzzy in the other. On the OSC4.4 this is P2’s job, played off against P1.
• Brightness / intensity is grid bias (the Z-axis baseline). Run it no brighter than you need: excess beam current burns the phosphor and softens focus (see burn/persistence in Vol 2).
• Centring is a DC offset on each plate pair (P6/P7). It moves the picture without resizing it.
• Size / deflection factor is the gain of each deflection amp (P4/P5). It trades picture size against the supply headroom of the amplifier.
• Geometry / linearity — pincushion, trapezoid, or non-linear spacing — is mostly fixed by the tube and the symmetry of the differential deflection drive (Vol 4). If a built clock shows geometry error, suspect an asymmetric plate drive or a marginal HV rail before you blame the firmware.

12.3 Troubleshooting

Work the table top-down. The cardinal rule of scope-clock debugging is the same as the bring-up philosophy: prove the spot exists before you blame the picture, and prove the voltages before you blame the spot. Discharge and verify (12.1.2) before probing anything with power removed; observe the one-hand rule (12.1.3) when probing live.

12.3.1 Fault → likely cause → check

Symptom	Likely cause	What to check
No spot at all	HV rail dead; heater not lit; grid bias too negative; broken/missing bring-up jumper	Confirm ~300 V at C6−/R8+ and the neon HV indicator lit; look for filament glow in the neck; verify G is not driven excessively negative (P3); confirm the U1 pin28→20 and pin8→26 jumpers are seated if testing pre-IC
Spot but no clock face	ICs missing or mis-ordered; MCU or DAC dead; wrong firmware	Re-run the staged insertion (12.2.2) — 4132s in U5/U4, 12F629 in U6, 7528 in U3, 18F26K20 in U1, all notch-correct; press S2; suspect MCU/DAC if a centred spot never becomes vectors
Defocused / fuzzy	Focus mis-set; astigmatism; brightness too high; weak/contaminated cathode	Re-converge P1 then P2 on a static spot; back off P3; if it will not sharpen at any setting, suspect the tube or a soft HV rail
Off-centre	Centring offset; unequal plate drive; deflection-amp imbalance	P6/P7 to recentre; if it will not centre, check both halves of the affected plate pair for equal swing (Vol 4)
Too dim	Grid too negative; low anode/accel voltage; tired phosphor/cathode	Raise P3 within reason; verify anode/plate rail at 150–250 V and the HV rail near 300 V; persistent dimness on an old tube is end-of-life
Too bright / blooming	Grid not negative enough; excess beam current	Lower P3; note that chronic over-brightness burns phosphor and softens focus
Jitter / waver	Supply ripple; poor grounding; ground loop	Scope the rails for ripple; tighten the star ground; verify reservoir caps (C1/C4/C6) are healthy and oriented correctly
Flicker	Refresh too low — too many objects drawn per frame	Reduce the number of drawn vectors/objects; a vector display flickers when the firmware cannot repaint the whole face fast enough (Vol 5)
One axis dead (line instead of a face)	Deflection amp or plate wiring on one axis failed	Check the dead axis’s amp stage and the X1/X2 or Y1/Y2 wiring to the tube; a single open plate collapses the picture to a line on the live axis

12.3.2 Diagnostic logic

The table compresses a simple decision tree. No spot is almost always power — HV, heater, or a grid biased into cutoff — so prove the test-point voltages in 12.5 first. Spot but no face is almost always the digital chain — wrong, missing, or mis-ordered ICs, or bad firmware — so re-run the staged insertion. Wrong-looking face is analog trim (focus, centring, size). Unstable face is the power/ground integrity of the supply (ripple, loops) or, for flicker specifically, the firmware drawing more than it can refresh. Half a face (a line) is a dead deflection axis. Each branch points you back at the volume that owns that subsystem.

12.4 (reserved — see the test points in 12.5)

The voltage test-point table is given its own section so it sits next to the cheatsheet for quick reference.

12.5 Voltage test-point table

All voltages are DC referenced to ground unless noted; the tab of the 7805 regulator is a convenient, reliable ground contact on the OSC4.4. Filament voltage is AC, measured across the two filament pins. These figures are the OSC4.4 measured values and are approximate but should be close; use the one nearest your tube.

Node	Symbol	Expected	Notes
Cathode	K	≈ −295 V DC	Beam accelerates from this negative cathode
Control grid (Wehnelt)	G	≈ −300 V DC	Grid bias; gates beam current (brightness)
Focus electrode	F	≈ −150 V DC	Set by P1 (focus)
X / Y plates + anode	X1, X2, A, Y1, Y2	≈ 150–250 V DC	Deflection plates and accelerating anode
Filament / heater	f — f	≈ 4–6 V AC (across the pair)	Set by Rfilament; verify against your tube’s spec on first power-up
Main HV rail	C6− / R8+	≈ 300 V DC	C6 negative, R8 positive; neon HV indicator should be lit
HV at grid node	G / C6−	≈ 300 V DC	G negative relative to C6−
Low-voltage logic rail	C44	+5 V DC	Left side +5 V, right side ground; LED lights when present

CRT pin-to-board mapping (the two tubes the OSC4.4 documents):

Function	DG7-32 pin	6Lo1i pin
Filament	1, 12	1, 14
Grid (G)	2	3
Cathode (K)	3	2
Focus (F)	4	4
Anode (A)	8	9
X1 / X2	10 / 9	8 / 7
Y1 / Y2	7 / 6	10 / 11

Resistor values track the tube: for the DG7-32, R99 & R4 = 330 K, R6 & R8 = 220 K; for the 6Lo1i, R99 & R4 = 220 K, R6 & R8 = 180 K. A 2BP1 uses R99 & R4 = 220 K. Filament dropping resistor Rfilament starts at 4.7 Ω 3 W for most tubes; some European types (e.g. a DG7-6) want 10 Ω 3 W. Always measure the filament voltage and adjust Rfilament up if it reads high.

12.6 Cheatsheet (laminate-ready)

Print, laminate, and keep on the bench. Everything below is self-contained.

12.6.1 Safe-discharge — every time, before touching

1. Power off, unplug at the wall.
2. Discharge each HV node through a bleeder/discharge resistor (1–10 MΩ, ≥2 W) with the other end on chassis ground / the CRT dag — never a bare screwdriver.
3. Measure 0 V with an HV-rated DMM. No 0 V reading = not safe.
4. If you walked away, discharge and re-measure (dielectric absorption recovers charge).
5. Live probing: one hand behind your back, isolation transformer in line, eye protection on if the tube is exposed, never alone on a kV supply.

12.6.2 Pot / control quick reference (OSC4.4)

Control	Function	Set for
P1	Focus	Roundest spot
P2	Anode / astig	Sharpest spot
P3	Brightness (grid)	As dim as readable
P4, P5	Image size X / Y	Picture fits, square
P6	X centre	Image centred L/R
P7	Y centre	Image centred U/D
R8 (Dutchtronix)	Image size on host scope	CCW = max, then fit

Preset for bring-up: P1–P5 mid, P7 fully CCW, P6 fully CW; jumper U1 28→20 and 8→26 for the blob. IC order: 4132→U5/U4, 12F629→U6, 7528→U3, 18F26K20→U1; press S2 for the face.

12.6.3 Voltage table (ref. 7805 tab = ground)

Node	Expected
K (cathode)	≈ −295 V DC
G (grid)	≈ −300 V DC
F (focus)	≈ −150 V DC
X1/X2/A/Y1/Y2	≈ 150–250 V DC
Filament f–f	≈ 4–6 V AC
HV rail (C6−/R8+)	≈ 300 V DC
Logic rail (C44)	+5 V DC

12.6.4 Fault → cause quick table

Symptom	First suspect
No spot	HV / heater / grid cutoff / missing jumper
Spot, no face	IC order / MCU / DAC / firmware
Fuzzy	Refocus P1 then P2; lower P3; tired tube
Off-centre	P6 / P7; unequal plate drive
Too dim	Raise P3; check anode & HV rails
Too bright	Lower P3 (burns phosphor)
Jitter / waver	Ripple, grounding, reservoir caps
Flicker	Too many objects — refresh too low
Half a face (line)	Dead deflection axis / plate wiring

12.7 Glossary

An A–Z of the terms used across this series.

• Anode (A1/A2). The accelerating electrode(s) that pull the beam toward the screen; A2 is the final/high-voltage anode, often tied to the aquadag. (Vol 2.)
• Aquadag. Conductive graphite coating on the inside (and a matching ground coat outside) of the tube; the inner coat is the final anode and forms a capacitor with the outer that stores charge after power-off. (Vol 2, 12.1.)
• Astigmatism. A spot that focuses sharp in one axis and fuzzy in the other; corrected by the astigmatism / spot-roundness control (OSC4.4 P2). (12.2.)
• Blanking. Cutting the beam off (via grid/cathode) during retrace so the connecting lines between strokes are invisible — the Z-axis doing its job. (Vol 4.)
• Cockcroft–Walton multiplier. A diode-and-capacitor voltage-multiplier ladder used in some designs to generate the high anode voltage from a lower AC source. (Vol 3.)
• Deflection factor. Volts-per-division of beam deflection — how much plate voltage moves the spot a given distance; the inverse of deflection sensitivity. (Vol 2, 4.)
• Electrostatic deflection. Steering the beam with voltages on internal plates (as in scope tubes), not coils — the architecture this whole series uses, as opposed to TV-style magnetic yokes. (Vol 2.)
• Getter. The silvery flash inside the envelope; a reactive metal that absorbs residual gas to maintain the vacuum. A white getter means air has leaked in and the tube is dead. (Vol 6.)
• HC. Horizontal-centring (and by analogy vertical-centring) offset — the DC bias that positions the picture; OSC4.4 P6/P7. (12.2.)
• Lissajous figure. The curve traced when one sinusoid drives X and another drives Y; the classic scope demonstration and the conceptual ancestor of vector drawing. (Vol 1, 2.)
• PDA (post-deflection acceleration). A high-voltage electrode after the deflection plates that boosts brightness without reducing deflection sensitivity; a major HV hazard in tubes that use it. (Vol 2, 12.1.)
• Persistence. How long the phosphor keeps glowing after the beam passes; short persistence flickers if refresh is low, long persistence smears motion. (Vol 2.)
• Phosphor (P1 / P7 / P31). Screen coating type: P1 is green medium-persistence (classic scope), P7 is the long-persistence blue/yellow two-layer (radar), P31 is bright green short-persistence. Not to be confused with pot designators P1–P7. (Vol 2, 6.)
• Raster. A display painted by scanning the whole screen in lines (TV); contrast with vector. Scope clocks are vector, not raster. (Vol 1, 2.)
• Vector display. A display that draws only the line segments and points it is told to, going directly from stroke to stroke — the scope-clock approach. (Vol 1, 2, 5.)
• Wehnelt / control grid (G). The cylinder around the cathode whose negative bias controls beam current (brightness) and, near cutoff, blanks the beam. (Vol 2, 4.)
• Z-axis. The brightness/blanking dimension (after the X and Y position axes); driven by the video amplifier into the grid or cathode. (Vol 4.)

12.8 References (Vol 12)

• Clocks hub shared safety baseline, _shared/safety.md (hazard tiers and the rules common to every high-voltage clock build).
• OSC4.4 assembly instructions — voltages, pot functions, CRT pinouts (DG7-32, 6Lo1i), and the staged bring-up/calibration sequence. Held in 02-inputs/OSC4_4 (I have this)/AssemblyInstructionsForOSC4.4.txt.
• Dutchtronix AVR Scope Clock — operating instructions (host-scope X-Y setup, R8 image-size control, S1/S2 menu and time/date set). Held in 02-inputs/A - Dutchtronix AVR Scope Clock (I have this)/AVR Scope Clock Operating.pdf.
• Cross-references within this series: Vol 2 (CRT physics, phosphors, vector vs raster), Vol 3 (HV supply generation and taming), Vol 4 (deflection and video/Z amplifiers), Vol 5 (timebase, DACs, vector drawing and refresh), Vol 6 (CRT selection and the getter), Vols 7 and 8 (the owned Dutchtronix and OSC4.4 builds), Vol 11 (enclosure creepage/clearance).
• E. Schlaepfer (TubeTime), “Electrostatic CRT Driver Design,” tubetime.us/?p=183 (HV and deflection design context for the open-source path).